- Aging- a:rd Service Wear of Check Valves Used---in Engineered · 2012. 11. 19. · NUREG/CR-4302...

NUREG/CR-4302ORNL-6193Vol. 2

- Aging- a:rd Service Wear ofi

I

Check Valves Used---in EngineeredSafety-Feature Systems of-Nuclear Power Plants

i.I

Aging Assessments and MonitoringMethod Evaluations

Prepared by

Prepared by 'H. D. Haynes

Oak Ridge National Laboratory

Prepared forU.S. Nuclear Regulatory Commission

I

AVAILABILITY NOTICE

< .- - Availability of Reference Materials Cited in NRC Publications -

Most documents cited in NRC publications will be available from one of the following sources:

1. .me NRC Public Document Room. 2120 L Street. NW. Lower Level. Washington, DC' 20555

2. -'The Superintendent of Documents. U.S. Government Printing Office. P.0. Box 37082, Washington.DC 20013-7082 ' - - ,

3. . The National Technical Information Service. Springfield, VA 22161

Although the listing that follows represents the majority of documents cited In NRC publications. It Is notIntended to be exhaustive.'

Referenced documents available for Inspection and copying for a fee from the NRC Public Document Roominclude NRC correspondence and internal NRC memoranda: NRC Office of Inspection and Enforcementbulletins. circulars. Information notices, Inspection and investigation notices: LUcensee Event Reports; ven-dor reports and correspondence: Cominssion papers: and applicant and licensee documents and corre-spondence.

The following documents In the NUREG series are available for purchase from the GPO Sales Program:formal NRC staff and contractor reports. NRC-sponsored conference proceedings, and NRC booklets andbrochures. Also avalable are Regulatory Guldes. NRC regulations In the Code of Federal Regulations, andNuclear Regulatory Commission Issuances.

Documents available from the National Tochrical Information Service Include NUREG series reports andtechnical reports prepared by other federal agencies and reports prepared by the Atomic Energy Commis-slon, forerunner agency to the Nuclear Regulatory Commission.

Documents available from' public and special technical libraries Include all open Iterature Items, such asbooks. journal and periodical articles, and transactions.. Federal Register notices, federal and state leglsla-tlon.oand congressional reports can usually be obtained from these libraries. .- '

Documents suchr as theses.- dissertations, foreign reports and translations, and non-NRC conference pro-ceedings are available for purchase from the organIzation sponsoring the publicatIon cited ;

Single copies of NRC draft reports are available free, to the extent of supply, upon written request to theOffice of Information Resources Management. Distribution Section. U.S. Nuclear Regulatory ComrmissIon. -

Washington. DC 20555.

Copies of indus*ix codes and standards used In a substantive manner In the NRC regulatory process aremaintained at the NRC Library. 7920 Norfolk Avenue, Bethesda, Maryland, and are available there for refer-ence use by the public. Codes and standards are usually copyrighted and may be purchased from theoriginating organization or, If they are American National Standards, from the American National StandardsInstitute, 1430 Broadway. New York, NY. 10018. - 4

r

I .

-'I - DISCLAIMER NOTICE '

This report was prepared as an account of work sponsored by an agency of the United States GovernmentNeitherthe United States Govemrnent norany agency thereof, or any oftheir employees, makes anywarranty,expresed or implied, or assumes any legal liability of responsibility for any third party's use. or the results ofsuch use, of any Information, apparatus, product or process disclosed in this report, or represents that Its use-by such third party would not Infringe privately owned rights. - - * - '

7

. - . 7.1.

NUREG/CR-4302ORNL 6193Vol. 2RV

Aging and Service Wear ofCheck Valves Used in EngineeredSafety-Feature Systems ofNuclear Power PlantsAging Assessments and MonitoringMethod Evaluations

Manuscript Completed: September 1990Date Published: April 1991

Prepared byH. D. Haynes

Oak Ridge National LaboratoryOperated by Martin Marietta Energy Systems, Inc.

Oak Ridge National LaboratoryOak Ridge, TN 37831-6285

Prepared forDivision of EngineeringOffice of Nuclear Regulatory ResearchU.S. Nuclear Regulatory CommissionWashington, DC 20555NRC FIN B0828Under Contract No. DE-ACO5-84OR21400

ABSTRACT

Check valves are used extensively in nuclear power plant safety systems andbalance-of-plant systems. The failures of these valves have resulted in significantmaintenance efforts and, on occasion,haveresulted in water hammer, overpressurization oflow-pressure systems, and damage to flow system components. These failures have largelybeen attributed to severe degradation of internal parts (e.g., hinge pins, hinge arms, discs,and disc nut pins) resulting from instability (flutter) of check valve discs under normal plantoperating conditions. Present surveillance requirements for nuclear power plant checkvalves have been inadequate for timely detection and trending of such degradation becauseneither the flutter nor the resulting wear can be detected prior to failure. Consequently,the U.S. Nuclear Regulatory Commission has had a continuing strong interest in resolvingcheck valve problems.

In support of the Nuclear Plant Aging Research Program, Oak Ridge NationpiLaboratory has carried out an evaluation of several developmental and/or commerciallyavailable check valve diagnostic monitoring methods, in particular, those based onmeasurements of acoustic emission, ultrasonics, and magnetic flux. In each case, theevaluations have been focused on the capability of each method to provide diagnosticinformation useful in determining check valve aging and service wear effects (degradation),check valve failures, and undesirable operating modes.

A description of each monitoring method is provided in this report, includingexamples of test data acquired under controlled laboratory conditions. In some cases, fieldtest data acquired in situ are also presented. The methods are compared, and suggestedareas in need of further development are identified.

iii

CONTENTS

LIST OF FIGURES ................. vii

LIST OF TABLES ................. xi

ACKNOWLEDGMENTS ................. 'axiii

SUMMARY ...................... xv

1. INTRODUCTION .... ... 1....1.1 NUCLEAR PLANT AGING RESEARCH GOALS .............. 11.2 NPAR PROGRAM STRATEGY ..... ................ 113 NRC INTEREST IN CHECK VALVES ..... .............. 2

1.3.1 Check Valve Function and Types ..................... 41.3.2 Selected NRC Notices and Bulletins ................... 51.3.3 INPO SOER 86-03 ............................... 61.3.4 Generic Letter 89-04 ........... .................... 61.3.5 EPRI Application Guidelines ........................ 7

1.4 PHASE I REPORT SUMMARY ........................... 71.4.1 General Information .............................. 71.4.2 Check Valve Failure Modes and Sites of Degradation ....... 81.4.3 Surveillance Requirements and Measurable Parameters ...... 81.4.4 Conclusions and Recommendations .................... 9

2. EVALUATION OF CHECK VALVE MONITORING METHODS ........ 112.1 GENERAL INFORMATION ............................. 112.2 ACOUSTIC EMISSION MONITORING ....................

2.2.1 Basic Principles ...............................2.2.2 Detection of Valve Disc Movement .................2.2.3 Qualitative Leak Detection .......................

2.3 ULTRASONIC INSPECTION ...........................2.3.1 Basic Principles ...............................2.3.2 Detection of Valve Disc Movement .................

2.4 MAGNE-TlC FLUX MONITORING .......................2.4.1 Basic Principles ...............................2.4.2 Detection of Valve Disc Movement ..................2.4.3 Detection of Worn Hinge Pins .....................

2.5 COMPARISON OF ACOUSTIC, ULTRASONIC, AND MAGNETICMETHODS ......................................

2.6 OTHER METHODS ..................................2.6.1 Radiography. .................................2.6.2 Pressure Noise Monitoring ........................

2.7 CONCLUSIONS-BENEFITS AND WEAKNESSES OFMONITORING METHODS ..........................

1111111617171719191921

23282828

31

v

3. DESCRIPTION OF SELECTED COMMERCIALLY AVAILABLECHECK VALVE DIAGNOSTIC SYSTEMS .......... ............ 333.1 CHECKMATED II (HENZE-MOVATS, INC.) .... ............ 333.2 QUICKCHECKm (LIBERTY TECHNOLOGY CENTER, INC.) .... 3333 VIP (CANUS CORPORATION) .......................... 353.4 AVLD (LEAK DETECTION SERVICES, INC.) .... ........... 363.5 MINAC-6 (SCHONBERG RADIATION CORPORATION) ....... 373.6 PHYSICAL ACOUSTICS CORPORATION .... .............. 37

4. NUCLEAR INDUSTRY CHECK VALVE GROUP TESTOF CHECK VALVE DIAGNOSTIC SYSTEMS .... ............... 394.1 GENERAL INFORMATION ............................. 394.2 SELECTED TEST RESULTS ............................ 39

5. SUMMARY AND RECOMMENDATIONS ........................ 51

REFERENCES . .............................................. 53

Appendix A.- REGULATORY TEST REQUIREMENTS AND CURRENT TESTMETHODS ....................................... 55

vi

LIST OF FIGURES

Figure Page

1.1 NPAR approach . ...................................... 1

1.2 Close-up of the hinge mechanism of a 10-in. swing check valvethat failed to close. 3

1.3 Additional lateral disc movement observed on the same failedcheck valve whose severely worn. hinge mechanismis shown in Fig. 1.2 ..................................... 3

1.4 Typical swing check valve ................................. 4

2.1 Acoustic signal vs time for a 10-in. check valve testedby Duke Power Company in March 1984 .12

2.2 Acoustic emission equipment (schematic representation) usedby Duke Power Company in 1987 flow loop tests .13

2.3 Acoustic waveform for a check valve during unstable (tapping)and stable (flow noise only) operations ....................... 14

2.4 Time-of-arrival technique ................................. 14

2.5 Check valve closures with new and artificially wornhinge pins . ........................................... 15

2.6 Check valve closures with two artificial degradations .......... I .... 16

2.7 Ultrasonic signatures during stable and unstable check valveoperations .18

2.8 Comparison of an ultrasonic signature with that from an RVDTinstalled on the hinge pin of a check valve .......... ..... 18

2.9 Magnetic flux signature analysis (MFSA) principleof operation .20

2.10 Comparison of magnetic field strength and disc angularposition measurements ................................. 21

2.11 Check valve flow loop (photo) .............................. 22

vii

Figure Page

2.12 Check valve flow loop (schematic) .......................... 23

2.13 Use of MFSA to detect disc instability (flutter) ................. 24

2.14 Detecting worn hinge pins using MFSA ....................... 24

2.15 Magnetic flux time waveforms of a 3-in. swing check valvefor three hinge pin conditions: normal (top trace), mediumworn (middle trace), and severely worn (bottom trace).All signatures were acquired using identical internalmagnet and external gaussmeter locations .25

2.16 Magnetic flux and acoustic signatures for a check valveunder several simulated operational conditions .................. 27

2.17 Pressure probe specifications and points of installation .... ........ 29

2.18 Effect of flow rate on pressure noise spectra ................... 29

2.19 Effect of flow path on pressure noise spectra ................... 30

3.1 A simplified depiction of the CHECKMATET™ II system .... ....... 34

3.2 CHECKMATED and acoustic signatures for a check valveundergoing backstop tapping .34

3.3 CHECKMATE" and acoustic signatures for a check valveundergoing seat tapping .................................. 35

3.4 Simplified depiction of the QUICKCHECKTh system .... ......... 36

4.1 QUICKCHECK7' signals for a 12-in. tilting-disc checkvalve in "new" condition .41

4.2 QUICKCHECKJ@ signals for a 12-in. tilting-disc checkvalve with 30% hinge pin diameter reduction .42

4.3 QUICKCHECK1h signals for a 12-in. tilting-disc checkvalve in "new' condition-expanded scale to showseating transient .43

4.4 QUICKCHECKam signals for a 12-in. tilting-disc checkvalve with 30% hinge pin diameter reduction-expandedscale to show seating transient .44

viii

Figurc Page

4.5 VIP (CANUS Corporation) accelerometer traces fromtest 410 (24-in. tilting-disc check valve, full-flowconditions, induced flow turbulence, worn hinge pin) .... .......... 45

4.6 VIP (CANUS Corporation) accelerometer traces fromtest 420 (24-in. tilting-disc check valve, full-flowconditions, minimal flow turbulence, worn hinge pin) .... ......... 46

4.7 Two ultrasonic signatures, obtained using the CHECKMATEh IIsystem for the same valve (10-in. Velan swing check) underdifferent flow conditions: 2251 gal/min (top trace)and 1295 gal/min (bottom trace) ............................ 47

4.8 A CHECKMATE' II signature for a 12-in. Val-Matic tilting-disccheck valve from the full-open to full-closed positions .... ......... 48

4.9 CHECKMATEV II signature for a 12-in. Val-Matic tilting-disccheck valve having a stuck disc. The top trace was obtainedat no-flow conditions, while the bottom trace was obtained at2392 gal/min (more than would be sufficient to move the disc) .... ... 49

ix

LIST OF TABLES

Table Page

1.1 Titles of selected NRC/XE Information Notices ..... ............. 5

1.2 Titles of selected NRCAE Bulletins ....... .................. 6

13 Check valve sites susceptible to aging-related degradation .... ...... 8

2.1 Selected diagnostic capabilities and limitations of threecheck valve monitoring methods .26

4.1 Check valves tested during the NIC monitoring methodevaluation tests ................................... 40

4.2 Check valve degradations used in the NIC tests ...... ........... 40

4.3 Flow conditions and valve operational modes used in the NIC tests ... 41

)d

ACKNOWLEDGMENTS

The check valve Phase II studies could have been carried out only with the helpof many people, whom the author wishes to acknowledge.

Appreciation is expressed for the advice and guidance provided during this studyby D. M. Eissenberg and W. S. Farmer. The author also wishes to thank D. A. Casada forhis many contributions, including the preparation of a summary of check valve regulatorytest requirements that is included verbatim in this report (Appendix A).

The help provided in the experimental portion of this study by E. L. Biddle, Sr.,D. E. Hendrix, and B. K Stewart is greatly appreciated.

Much information was obtained from diagnostic system vendors in the form ofproduct literature and visits to their facilities. In addition, much was learned fromwitnessing one day of check valve tests at the Utah Water Research Laboratory. In thatregard, the author would like to thank the following people for their assistance: W. M.Suslick (Duke Power Company), D. M. Ciesielski and J. N. Nadeau (Henze-Movats, Inc.),P. Pomaranski (CANUS Corporation), D. Manin (Liberty Technology Center, Inc.), J. W.McElroy (Philadelphia Electric Company), and M. T. Robinson (Chairman-NuclearIndustry Check Valve Group).

The assistance of Jeanie Shover and Janice Asher in preparing this report isgratefully acknowledged.

xiii

SUMMARY

In support of the U.S. Nuclear Regulatory Commission's Nuclear Plant AgingResearch Program, Oak Ridge National Laboratory (ORNL) carried out an evaluation ofseveral developmental and/or commercially available check valve diagnostic monitoringmethods. Assessments were made of the capability of each method to provide diagnosticinformation useful in determining check valve aging and service wear effects (degradation),check valve failures, and undesirable operating modes. These methods included

* Acoustic emission monitoring.* Ultrasonic inspection.* Magnetic flux signature analysis.* Radiography.* Pressure noise signature analysis.

The evaluations have been focused on the capability of each method to providediagnostic information useful in determining check valve aging and service wear effects(degradation), check valve failures, and undesirable operating modes. Commercial suppliersof three check valve monitoring systems recently participated in a comprehensive series oftests designed to evaluate the capability of each monitoring technology to detect theposition, motion, and wear of check valve internals (e.g., disc, hinge arm, etc.) and valveseat leakage. These tests, directed by the Nuclear Industry Check Valve Group and carriedout at the Utah Water Research Laboratory, are described in this report.

Of those methods examined by ORNL, acoustic emission monitoring, ultrasonicinspection, and magnetic flux signature analysis provided the greatest level of diagnosticinformation. These three methods were shown to be useful in determining check valvecondition (e.g., disc position, disc motion, and seat leakage), although none of the methodswas, by itself, successful in monitoring all three condition indicators. However, thecombination of acoustic emission with either ultrasonic or magnetic flux monitoring yieldsa monitoring system that succeeds in providing the sensitivity to detect all major check valveoperating conditions. When monitoring check valves using either of these combinations,it is not necessary to apply both methods simultaneously in order to achieve the combinedcapabilities of the two methods. Commercial versions of all three methods are still underdevelopment, and all should improve as a result of further testing and evaluation.

xv

1. INTRODUCTION

1.1 NUCLEAR PLANT AGING RESEARCH GOALS

This document describes work performed in support of the U.S. Nuclear RegulatoryCommission's (NRC's) Nuclear Plant Aging Research (NPAR) Program, which wasestablished primarily as a means to resolve technical safety issues related to the aging ofelectrical and mechanical components, safety systems, support systems, and civil structuresused in commercial nuclear power plants.! A comprehensive Phase II aging assessment oncheck valves was performed by Oak Ridge National Laboratory (ORNL), the results ofwhich are' presented in this report.

The goals of the NPAR Program are as follows:

1. To identify and characterize aging effects that, if unchecked, could cause degradationof components, systems, and civil structures and thereby impair plant safety.

2. To identify methods of inspection, surveillance, and monitoring and of evaluating theresidual life of components, systems, and civil structures that will ensure timely detectionof significant aging effects before loss of safety function.

3. To evaluate the effectiveness of storage, maintenance, repair, and replacement practicesin mitigating the rate and extent of degradation caused by aging.

1.2 NPAR PROGRAM STRATEGY

The methodology employed by the NPAR Program is basically a two-phaseapproach, as illustrated by Fig. 1.1. This strategy is applicable for all components, systems,and structures selected for aging assessments. Research efforts are conducted in accordancewith the objectives associated with each phase.

ORNL-WSC-4060 ETD

Phase I

* Operating Experience Phase It Recommendations forSelect Review and Analysis * VerifIcation of Improved * Evaluation of

Systems C Review of Methods IS&MM* 'Mothballed" Equipmentand and Technology for 0. Tests of Naturally Aged * Modifcation of Codes.

Components IS&MM Components: aendtfor * Screening Type Components. Models, end StendLRds. nd Guides

Assessment Examinalion and Tests - Samples with Simulated * Licn Generic* Interim Recommendations Degradation

for Engineering Tastse Cost/lenefl tStudyIn Phase ii

*IS&MM - Inspection. surveillance, end monitoring methods

Fig. 1.1. NPAR approach.

l

2

The objectives for an NPAR Phase I aging assessment of a component, system, orstructure are as follows:

1. Identify and characterize aging and wear effects.2. Identify failure modes and causes attributable to aging.3. Identify measurable performance parameters, including functional indicators.

Phase I studies result in a determination of whether additional research is needed;if so, recommendations are developed to identify and guide further studies.

An NPAR Phase II assessment is carried out with the following objectives:

1. Perform in-depth engineering studies and aging assessments based on in situmeasurements.

2. Identify improved methods for inspection, surveillance, and monitoring or for evaluatingresidual life.

3. Perform postservice examinations and tests of naturally aged/degraded components.4. Make recommendations for utilizing research results in the regulatory process.

The results of a Phase II aging assessment are intended to form the basis forimplementing improved inspection, surveillance, maintenance, and monitoring methods;modifying present codes and standards; developing guidelines and review procedures forplant life extension; and resolving generic safety issues.

13 NRC INTEREST IN CHECK VALVES

Check valves are used extensively in nuclear plant safety systems andbalance-of-plant (BOP) systems. The failures of these valves have resulted in significantmaintenance efforts and, on occasion, have resulted in water hammer, overpressurizationof low-pressure systems, and damage to flow system components. These failures havelargely been attributed to severe degradation of internal parts (e.g., hinge pins, hinge arms,discs, and disc nut pins) resulting from instability (flutter) of check valve discs under normalplant operating conditions. For example, a post-service examination was carried out on a10-in. swing check valve from a local installation following its failure to close in service.The cause of failure was determined to be extreme wear in the hinge mechanism, whichpermitted the disc to hang up on the valve body before seating occurred. Although serviceconditions are not, known, the wear appears to have been a result of disc instability(oscillation) during service. A close-up of the hinge mechanism for this valve, which showsthe severe wear observed and its effect on disc position, is shown in Figs. 1.2 and 1.3,respectively. Check valve instability may be a result of misapplication (using oversizedvalves) and may be exacerbated by low-flow conditions and/or upstream flow disturbances.2

Present surveillance requirements for nuclear power plant check valves have beeninadequate for timely detection and trending of such degradation because neither the flutternor the resulting wear can be detected prior to failure. Consequently, the NRC has hada continuing strong interest in resolving check valve problems.

I

3

ORNL PHOTO 7382-90

_ I,I

- i

I

Fig. 12 CQosc-up of the hinge mechanism of a 10-in. swing check valve that failedto close.

ORNL PHOTO 7998-86, .. _~~~~~~-

Fig. 1.3. Additional lateral disc movement observed on the same failed check valvewhose severely worn hinge mechanism is shown in Fig. 1.

4

13.1 Check Valve Function and Types

The function of a check valve is simply to open and thus permit flow in only onedirection. When the flow stops or reverses direction, the check valve closes. Check valvesare self-actuating; that is, they require no external mechanical or electrical signal to eitheropen or close. As a result, most check valves are not capable of being actuated other thanby changing the flow through the valve. Several types of check valves are commonly used,such as the swing check, piston-lift, ball, stop-check, and duo-check designs. Thedescriptions of check valve monitoring methods in this report refer in most cases to theiruse on the swing check valve, shown in Fig. 1.4. However, all monitoring methodsdescribed herein have the potential for being applied to other check valve types. A morecomprehensive description of check valve types appears in Ref 3.

ORNL-DWG 884895A ETD

Fg. 1.4. Typical swing check valve.

5

13.2 Selected NRC Notices and Bulletins

Tables 1.1 and 1.2 contain the titles of selected NRC Inspection and Enforcement(IE) Notices and Bulletins over the last 10 years; these notices and bulletii's are indicativeof the types of check valve problems that have occurred during this period and of thecontinued interest that the NRC has had in identifying and solving these problems.

Table 1.1. Titles of selected NRC/HE Information Notices?

Number Title

90-03 Malfunction of Borg-Warner Bolted Bonnet Check Valves Caused byFailure of the Swing Arm

89-62 Malfunction of Borg-Warner Pressure Seal Bonnet Check Valves Causedby Vertical Misalignment of Disc

88-85 Broken Retaining Block Studs on Anchor Darling Check Valves

88-70 Check Valve In-Service Testing Program

86-06 Failure of Check and Stop Check Valves Subjected to Low-FlowConditions

86-01 Failure of Main Feedwater Check Valves Causes Loss of FeedwaterSystem Integrity and Water Hammer Damage

84-12 Failure of Soft Seat Valve Seals

84-06 Steam Binding of Auxiliary Feedwater Pumps

83-54 Common Mode Failure of Main Steam Isolation Nonreturn Check Valves

83-06 Nonidentical Replacement Parts

82-35 Failure of Three Check Valves on High-Pressure Injection Lines to PassFlow

82-26 Reactor Core Isolation Cooling and High-Pressure Coolant InjectionTurbine Exhaust Check Valve Failures

82-20 Check Valve Problems

82-08 Check Valve Failures on Diesel Generator Engine Cooling System

81-35 Check Valve Failures

81-30 Velan Swing Check Valves

80-41 Failure of Swing Check Valve in the Decay Heat Removal System atDavis-Besse Unit No. 1 -

aThese information notices are available in the NRC Public Document Room.

6

Table 12. Titles of selected NRCAIE Bulletinsa

Number Title

89-02 Stress corrosion cracking of high-hardness type 410 stainless steel internalpreloaded bolting in Anchor Darling Model S35OW swing check valves orvalves of similar design

85-01 Steam binding of auxiliary feedwater pumps

83-03 Check valve failures in raw water cooling systems of diesel generators

80-01 Operability of automatic depressurization system (ADS) valve pneumaticsupply

aThese bulletins are available in the NRC Public Document Room.

In particular, IE information Notice 86-01 describes an event that occurred at SanOnofre, Unit 1, on November 21, 1985. The most significant aspect of the event was thefailure of five safety-related feedwater system check valves (three main feedwater regulatorcheck valves and two feedwater pump discharge check valves). The failure of these valveswas the primary cause of a severe water hammer that extensively damaged a portion ofthe feedwater system. The details of this event have been described in NUREG-1 190, Lossof Power and Water Hammer Event at San Onofre, Unit 1, on November 21, 1985.4 Thisreport presents the findings and conclusions of an NRC Incident Investigation Team sentto San Onofre by the NRC Executive Director for Operations.

133 INPO SOER 86-03

In 1986, the Institute of Nuclear Power Operations (INPO) issued a significantoperating experience report (SOER), numbered 86-03, which recognized the check valveproblems facing the nuclear industry and recommended that nuclear power plants establisha preventive maintenance program to ensure check valve reliability. INPO furtherrecommended that the maintenance program should include periodic testing, surveillancemonitoring, and/or disassembly and inspection.

13.4 Generic Letter 89-04

In April 1989, the NRC issued Generic Letter (GL) 89-04 (Ref. 5) in recognitionof the differences among utilities in the scope of valves included in in-service test (IST)programs and concerns about methods of fulfilling the requirements of 10 CFR 50.55a(g),which requires that certain pumps and valves be tested to assess their operational readinessin accordance with the Sect. XI requirements of the American Society of MechanicalEngineers (ASME) Boiler and Pressure Vessel Code. GL 89-04 describes potential genericdeficiencies associated with full-flow testing and back-flow testing of valves and analternative to full-flow testing (disassembly and inspection). It should be noted that GL89-04 addresses other aspects of IST programs as well.

At present, the nuclear industry, led by the Nuclear Industry Check Valve Group(NIC), is preparing guidelines for exemptions to GL 89-04, including methodologies for

7

extending the disassembly and inspection interval defined by GL 89-04 and guidelines foralternatives to full-flow and back-flow testing of check valves.

135 EPRI Application Guidelines

In January 1988, EPRI issued a set of guidelines for the selection, installation, andmaintenance of check valves in nuclear power plants.2 The stated major objective of thisdocument is to provide accurate technical information to be used by utilities in assessingthe long-term reliability of their check valve installations and to assist them in respondingto the SOER 86-03. It is noted that this information can also be used in selecting checkvalves for new systems and for plant modifications. The document contains considerabletechnical information and includes descriptions of several monitoring methods that wereavailable at the time the guidelines were prepared. The monitoring methods that aredescribed are as follows:

* Noise/Vibration Monitoring.* Acoustic Emission Monitoring.* Ultrasonic Methods.* Radiography.* Fiber Optics.

An evaluation of four of these monitoring methods is included in this report. Anassessment of fiber optics inspection methods was not carried out.

1.4 PHASE I REPORT SUMMARY

1.4.1 General Information

ORNL has completed a Phase I aging assessment on check valves.3 Major topicscovered by the study included

1. Check valve design features.2. Plant operating experiences.3. Surveillance requirements.4. Maintenance practices.5. Failure modes and causes.6. Parameters to monitor.

Results from this study were based primarily on information from operatingexperience records, including the Licensee Event Report (LER) file, the Nuclear PlantReliability Data System (NPRDS), and the In-Plant Reliability Data System (IPRDS). Inaddition to these data bases, information was gathered from component manufacturers byreviewing their literature and participating in discussions with their representatives.

8

1.4.2 Check Valve Failure Modes and Sites of Degradation

Five check valve failure modes were identified by the Phase I study:

1.2.3.4.5.

Failure to open.Failure to close.Plugged (limited or no flow through a normally open valve).Reverse (internal) leakage.External leakage.

Several check valve sites were identified as being susceptible to aging-relateddegradation. These sites and the corresponding aging mechanisms are presented in Table1.3.

1.43 Surveillance Requirements and Measurable Parameters

Test requirements for nuclear plant check valves were covered briefly in the Phase Ireport' and are summarized in this section. A more detailed description of the regulatoryrequirements related to check valve testing is presented in Appendix A of this report.

Testing requirements for nuclear plant check valves are contained in the plantTechnical Specifications and are in accordance with Sect. XI of the ASME Boiler and

Table 13. Check valve sites susceptible to aging-related degradation

Site Aging mechanism

Body assembly Body wear, erosion, corrosion

Fastener loosening, breakage

Internals Hinge pin wear, erosion,corrosion

Hinge pin fracture

Hinge arm wear, fracture

Disc nut loosening, tightening

Disc nut breakage

Disc wear, erosion, corrosion

Seat wear, erosion, corrosion

Foreign material

Seals Cap gasket deterioration

9

Pressure Vessel Code. Article IWV-3000 of the ASME Code describes in-service inspectionrequirements for check valves. This requirement consists primarily of exercising the valveto verify obturator (e.g., disc) travel to or from the full-open and full-closed positions asrequired to fulfill the valve's isolation function. Confirmation of obturator movement maybe by visual observation, a position indicator, observation of relevant pressures in thesystem, or other positive means.

Some check valves used for containment isolation are also required to be testedin accordance with 10 CFR 50, Appendix J. These tests involve pressurizing the checkvalve locally in the same direction as when the valve is required to perform its safetyfunction and comparing leakage rates through the valve with the specified standard.

These tests are intended to demonstrate check valve operability under test conditionsbut do not ensure check valve actuation as required under other anticipated operatingconditions. In addition, these tests are inadequate for timely detection and trending ofcheck valve degradation.

Check valve measurable parameters identified in the Phase I report3 as importantfor evaluating operational readiness include force or torque applied to move the obturator;fluid level, temperature, pressure, pressure differential, and flow rate; reverse leakage rate;humidity; and noise. Additional parameters identified as being necessary for positive failure-cause identification and for enhancement of capabilities for degradation tracking andincipient failure detection include dimensions, appearance, roughness, cracking, and bolttorque.

1.4.4 Conclusions and Recommendations

Three major methods that are used for check valve failure-cause identification wereidentified in the Phase I report.? These methods are external visual examination, inspectionduring normal maintenance, and periodic disassembly and inspection. Potentially usefulmeasurable parameters for detection of degradation and incipient failure were identified.The effectiveness and acceptability of these parameters were to be determined by furtherstudy (e.g., the Phase II aging assessment).

11

2. EVALUATION OF CHECK VALVE MONITORING METHODS

2.1 GENERAL INFORMATION

The primary objective of the Phase II check valve aging assessment program is toidentify and recommend methods of inspection, surveillance, and monitoring that wouldprovide timely detection of check valve degradation and service wear (aging) so thatmaintenance or replacement could be performed prior to loss of safety function(s). In thatregard, ORNL has carried out an evaluation of several developmental and/or commerciallyavailable check valve diagnostic monitoring methods, in particular, those based onmeasurements of acoustic emission, ultrasonics, and magnetic flux. The evaluations havebeen focused on the capability of each method to provide diagnostic information useful indetermining check valve aging and service wear effects (degradation), check valve failures,and undesirable operating modes.

A description of each monitoring method, including examples of test data acquiredunder controlled laboratory conditions, is provided in this report. In some cases, field testdata acquired in situ are also presented. The methods are compared, and suggested areasin need of further development are identified.

2.2 ACOUSTIC EMISSION MON1TORING

22.1 Basic Principles

Acoustic emissions (pressure waves) can be generated in a variety of ways. Ofparticular interest are those generated either when solids contact each other or when liquidsor gases flow through pipes and fittings., Acoustic emissions are detected by sensors, suchas piezoelectric-type accelerometers or microphones, which respond to pressure waves overa wide range of frequencies. Signal-conditioning electronics can be used to amplify selectedacoustic signals while attenuating others, e.g., unwanted environmental background noise.Analyses of acoustic emission signals obtained from check valves can be used to monitorcheck valve disc position, movement, and mechanical condition, as well as internalflow/leakage through the valve. Various methods of obtaining and analyzing data areavailable. These include filtering, fast Fourier transform (FF1), computer analysis, etc.

22-2 Detection of Valve Disc Movement

Acoustic emission monitoring has been shown to detect check valve disc movement.As an example, Duke Power Company (DPC) installed an acoustic sensor on top of a 10-in.cold-leg accumulator discharge check valve.*. A schematic representation of the installationis given in Fig. 2.1. After initially charging the accumulator to 100 psig, the motor-operated

*W. M. Suslick, DPC, "Proposed Technique for Monitoring Check ValvePerformance," presented at the INPO Check Valve Technical Workshop, October 30-31,1986.

12

ORNL-DWG 89-5100 ETD

SYSTEM SCHEMATIC

--- COLD LEG ACCUMULATOR C)

CCDi

MOTOR-OPERATED VALVE BYr (STROKE TIME -10 SECONDS) IN

v F MONITORED CHECK VALVE

rCLE TEST OF A TEN INCH)LD LEG ACCUMULATORSCHARGE CHECK VALVE'DUKE POWER COMPANYMARCH 1984

)M

ACOUSTIC EMISSION SIGNATURETHROTTLING NOISE

TIME 'i I.AC

I(S)

Fig. 2.1. Acoustic signal vs time for a 10-in, check valve tested by Duke PowerCompany in March 1984. Source: W. M. Suslick, "Proposed Technique for MonitoringCheck Valve Performance," presented at the INPO Check Valve Technical Workshop,October 30-31, 1986.

discharge valve was cycled. The acoustic sensor output during this cycling was processedand displayed on a strip chart recorder. The resulting acoustic signature (Fig. 2.1) showsthat the sensor detected the metal-to-metal contact occurring at the end of both theopening and closing strokes.

DPC has also carried out check valve acoustic emission testing under controlledflow loop conditions and with the introduction of various implanted defects that simulatedsevere aging and service wear.* Accelerometers were strapped to the bodies of threecheck valves in a manner depicted in Fig. 2.2. The following discussion summarizes theresults obtained from those tests.

*W. M. Suslick, H. F. Parker, and B. A. McDermott, DPC, "Acoustic EmissionMonitoring of Check Valve Performance," presented at the EPRI Power Plant ValvesSymposium, October 11-12, 1988, Charlotte, NC.

13


FffC - FFr OSCILLOSCOPE

Qn Q7 -7 CHANNEL TAPE RECORDER

- AUDIO SPEAKER

-PREAMP

- X CHECK VALVE

- ACCELEROMETER

Fig. 2.2 Acoustic emission equipment (schematic reprcsentation) used by DukePower Company in 1987 flow loop tests. Source. W. M. Suslick, H. F. Parker, and B. A.McDermott, "Acoustic Emission Monitoring of Check Valve Performance," presented at theEPRI Power Plant Valves Symposium, October 11-12, 1988, Charlotte, NC.

Tapping of the valve disc against its backstop was easily detected and distinguishedfrom background flow noise, as shown in Fig. 2.3. -In addition, by using two (or more)valve-mounted acoustic sensors, DPC was able to approximately locate the source of thetapping based on a comparison of the "time of arrival' of the acoustic signals acquired fromthe two sensors. An example of this technique is shown in Fig. 2.4.

By using the acoustic emission check valve monitoring method demonstrated byDPC, it appears likely that the following check valve operational conditions can bedetermined:

* Valve rapid opening (backseat impact).* Valve disc tapping during reduced flow. -

* Hinge arm tapping during reduced flow.* Valve rapid closing (seat impact).

Although a fully open check valve could be assumed by the existence of flow noise withoutthe presence of tapping, the absence of detectable tapping noise is itself no guarantee thatthe check valve is fully open since the valve disc may be -oscillating without tapping inmidstroke, may have fallen off, or may be stuck in a position that prevents it fromimpacting the valve body at any location.

Several tests were carried out by DPC on an 8-in. check valve in new conditionand with simulated degradation. Hinge pin diameters and disc/hinge arm clearances wereboth varied during valve cycle tests that generated acoustic emission signatures duringopening and closing.

14

-

-j

znU)

En

8e-


UNSTABLE OPERATION

= = I __ ISTABLE OPERATION _

. * I I l.T

, -I--- 1__1_ -

TIME

Fig. 2.3. Acoustic waveform for a check valve during unstable (tapping) and stable(flow noise only) operations. Source: W. M. Suslick, H. F. Parker, and B. A. McDermott,"Acoustic Emission Monitoring of Check Valve Performance," presented at the EPRI PowerPlant Valves Symposium, October 11-12, 1988, Charlotte, NC.

zC

)cd,-

8

uzS2

-ezQC

-Jc< 2

< 2

C,


== =E _ =~ II=

_ U~~~I

- -- V2 t1@-4@40 T T- I Iq77

- T134 LME

TAPPING SENSOR RESPONDINGLOCATION TO PRESSURE WAVE FIRST

HINGE PINBACKSTOP

THE ONE NEAREST THE HINGE PINTHE ONE NEAREST THE BACKSTOP

Fig. 2.4. Time-of-arrival technique. Source: W. M. Suslick, H. F. Parker, and B. A.

McDermott, "Acoustic Emission Monitoring of Check Valve Performance," presented atthe EPRI Power Plant Valves Symposium, October 11-12, 1988, Charlotte, NC.

15

Valve closures with new and artificially worn hinge pins are illustrated in Fig. 2.5;this figure shows that, with the worn hinge pins, an acoustic transient preceded the seatimpact. This transient may result from impact between the hinge pin and hinge armsurfaces as a result of the increased clearance between these two parts.

A similar transient event occurred as a result of increased clearance between thedisc stud and hinge arm, as illustrated in Fig. 2.6. Also shown is a closure of a check valvehaving both a worn hinge pin and a loose disc/hinge arm connection.

In practice, DPC has monitored check valve acoustic emission at three plants for- 1 year using valve-mounted accelerometers whose outputs are recorded on tape (at thecheck valves). Recorded signals are then played back remote from the valve (e.g., in anoffice) to both an oscilloscope and a loudspeaker for audio interpretation. Valve instabilityis then detected as tapping (clicking) noises heard on the loudspeaker and is quantified bythe oscilloscope.

DPC uses acoustic emission monitoring primarily to identify valves that are operatingin an unstable manner. Those valves are then targeted for close inspection (disassembly)


-j

0P:Wn

z

C,)0F:C',

-I

I.-Cn

NEWHINGE PIN

0.20-INCHUNDERSIZEDHINGE PIN

0.40-INCHUNDERSIZEDHINGE PIN

TIME (20 ms/div)

Fig. 2.5. Check valve closures with new and artificially worn hinge pins. Source:W. M. Suslick, H. F. Parker, and B. A McDermott, "Acoustic Emission Monitoring ofCheck Valve Performance," presented at the EPRI Power Plant Valves Symposium, October11-12, 1988, Charlotte, NC.

16


ACOUSTIC EMISSION WAVEFORM OF SWINGCHECK VALVE CLOSURE WITH WORN DISC STUD CONNECTION

-J

ZSo

(-'I,

I I I!,l1jeI I = 5if *I1 II l -tI lll I I

T I = =

TIME

ACOUSTIC EMISSION WAVEFORM OF SWING CHECKVALVE CLOSURE WITH WORN HINGE PIN AND DISC ARM CONNECTION

tIic IIII.1 1I1 1TnME

Fig. 26. Check valve closures with two artificial degradations. Source: W. M.Suslick, H. F. Parker, and B. A. McDermott, "Acoustic Emission Monitoring of Check ValvePerformance," presented at the EPRI Power Plant Valves Symposium, October 11-12, 1988,Charlotte, NC.

at the next convenient time. Roughly 10% of the valves monitored by DPC weredetermined to be operating in an unstable mode. After disassembly, roughly 90% of thesevalves showed signs of degradation (wear).

223 Qualitative Leak Detection

Acoustic emission techniques have long been used to detect fluid leaking througha valve. Philadelphia Electric Company (PECO) has been using acoustic techniques todetect valve leakage in their nuclear power plants since 1974.* Their test procedureconsists of acquiring two sets of valve acoustic emission readings, one while the valve isunpressurized and one with a pressure difference across the (closed) disc. The noiseassociated with a leaking valve is then determined on the basis of the difference in readings.

PECO has had good success with a portable, battery-powered data acquisition unitfor leakage monitoring. The acoustic data collected from baseline (unpressurized) andpressurized tests are downloaded into a computer for analysis, trending, and archiving.

*J. W. McElroy, PECO, "Light Water Reactor Valve Performance Surveys UtilizingAcoustic Techniques," presented at the EPRI Power Plant Valves Symposium, August25-26, 1987, Kansas City, MO.

17

23 ULTRASONIC INSPECTION

2.3.1 Basic Principles

Ultrasonic inspection involves the introduction of high-frequency sound waves intoa part being examined and an analysis of the characteristics of the reflected beam.

Typically, one (pulse-echo) or two (pitch-catch) ultrasonic transducers are used whichprovide both transmission and receiving (sensing) capabilities. The ultrasonic signal isinjected from outside the valve by the transmitting transducer and passes through the valvebody, where it is reflected by an internal part (e.g., disc, hinge arm, etc.) back toward thereceiving transducer. (Note: When one transducer is used in a pulse-echo mode, it providesboth transmitting and receiving capabilities.) By knowing the time required for transmissionof the ultrasonic signal from the transmitting transducer and back to the receivingtransducer, the transducer location(s), and other valve geometries, the instantaneous discposition may be determined.

In general, signal processing circuitry must be used to filter out undesirableultrasonic signal reflections present in the raw received signal so that the resultant processedsignal provides a more easily interpreted valve disc position signature. Various methods ofobtaining and analyzing data are available. These include filtering, FFT, computer analysis,etc.

23.2 Detection of Valve Disc Movement

Ultrasonic inspection techniques can be used to produce a time waveform displayfrom which disc position and movement may be easily determined. A properly conditionedultrasonic signal time waveform can be used to detect the following check valve operationalmodes:

Operational Mode Signature Characteristic

Full open or full closed Steady signal

Free flutter Variable signal with rounded peaks

Backstop tapping Similar to free flutter butwith flattened upper* peaks

Seat tapping Similar to free flutter butwith flattened lower* peaks

*Dependent on mounting position of the ultrasonic transducer(s).

Figure 2.7 shows ultrasonic signatures taken from a swing check valve installed ina laboratory flow loop at two disc positions: full open and partially open. It is noted thatthe disc was fluttering (unstable) in the partially open position and that the flutter wasclearly detected using the ultrasonic method. Figure 2.8 illustrates the similarity betweendisc motion signatures acquired with an ultrasonic sensor and with a specially installed rotaryvariable differential transformer (RVDT) attached directly to the hinge pin to provide adirect measurement of disc position.

18


0.21

0.00

5 -0.21

-0.41

-0.62 L

CHECKMATEV2\T6

DISC AT STABLE POSITION

DISC AT lUNSTABLE POSITION

CHECKMATE* J\ JSVZT5

0.91 r

0.60

B 0.300

0.00

-a.: If% . |Dv I

0.0 0.4 0.8 1.2 1.6 2.0TIME (s)

2.4 2.8 3.2 3.6 4.0

Fig. 2.7. Ultrasonic signatures during stable and unstable check valve operations.(Used with the permission of Henze-Movats, Inc.)


0.16 r

o.coQ)

.0.16

:O.. I , \~) 1 .,1( \../

CHECKMATE

,A /~ j . (

\J.11 \ ~/ \110.3 _-

-0.49 L_CHECKMATE VS RVDT ON HINGE PIN

0.21 r

ai O.c0-j

o -0.21

-0.41

RVDT

- N 4.4 *AJ *tx .j A VZ"N y rw .\y ,, / rkksd/"N\V ~,

-0.62 _-0.0

I I I I I I I I I I

0.4 0.8 1.2 1.6 2.0TIME (s)

2.4 2.8 3.2 3.6 4.0

Fig. 2.8. Comparison of an ultrasonic signature with that from an RVDT installedon the hinge pin of a check valve. (Used with the permission of Henze-Movats, Inc.)

19

Ultrasonic inspection techniques can be used to detect a missing or stuck disc. Forexample, if the disc is missing, no signal will be returned (reflected) from the disc; however,if the hinge arm remains on the valve, the hinge arm position can be verified by ultrasonicinspection techniques. Furthermore, under similar flow conditions, a hinge arm without anattached disc will flutter at higher frequencies than if a disc were attached.*

Disc stud wear can be detected by ultrasonic inspection by monitoring the motionof both the disc and hinge arm using ultrasonic transducers, one sensing movement of thedisc and the other sensing hinge arm movement. Increased clearance between the disc studand the hinge arm can result in increased movement of the disc relative to the hinge arm.

2.4 MAGNETIC FLUX MONITORING

2.4.1 Basic Principles

Research carried out by ORNL as part of the NPAR Phase II study of check valveshas led to the identification of a new check valve diagnostic technique, magnetic fluxsignature analysis (MFSA).t MFSA is based on correlating the magnetic field strengthvariations monitored on the outside of a check valve with the position of a permanentmagnet placed on a moving part inside the check valve (Fig. 2.9).

In the proof-of-principle tests, a Hall-effect gaussmeter probe was used outside thecheck valve to detect the magnitude of the magnetic field produced by a small cylindricalor rectangular (bar) permanent magnet attached to the hinge arm. The Hall-effect probedetected both constant and varying magnetic fields and thus continuously monitored boththe instantaneous position and the motion of the check valve disc. Various methods ofobtaining and analyzing data are available. These include filtering, FFT, computer analysis,etc.

2.4.2 Detection of Valve Disc Movement

MFSA provides the ability to monitor disc position through an entire valve strokeusing one externally mounted sensor. A comparison of disc position measured mechanically(by'an angular displacement transducer attached to the hinge pin) with that obtained byMFSA is shown in Fig. 2.10 for a 3-in. swing check valve whose disc was moved manually.MFSA has been applied to several swing check valves having different body materials andranging in size from 2 to 10 in.

MFSA also provides indication of disc flutter. This was demonstrated by' testscarried out by ORNL on a 2-in. swing check valve that was installed in a water flow loop.The ORNL check valve flow loop, illustrated photographically in Fig. 2.11 and schematicallyin Fig. 2.12, utilized a centrifugal pump that is capable of delivering >300 gal/min througha 2-in. nominal-diameter line (>30 ft/s).

*R. D. Ryan, CHECKMATE"' Il-A Diagnostic Tool for Check Valves, Henze-Movats, Inc., Technical Resources, May 1990.'

tH. D. Haynes and D. M. Eissenberg, ORNL, "Performance Monitoring of SwingCheck Valves Using Magnetic Flux Signature Analysis," Information Package ContainingSelected MFSA Test Results, May 1989.

20


HALL-EFFECTDSENSOR

MAGNET7

| I # I~~I

Fig. 2.9. Magnctic flux signature analysis (MFSA) principle of operation.

Flow rate through the tested check valve was controlled by means of three ballvalves located downstream of the check valve and a manually operated gate valve locatedin a short, 3-in. nominal-diameter bypass pipe circuit. Low flow rates through the checkvalve were achieved by fully opening the bypass gate. If needed, the flow through thecheck valve was further throttled by the ball valves. Conversely, higher flow rates throughthe check valve were achieved by throttling the gate valve.

The flow loop contains three flowmeters: a 24-gal/min rotameter, a 60-gal/minrotameter, and a turbine flowmeter. At flow rates less than approximately 20 gal/minthrough the check valve (<2 ft/s), the turbine flowmeter does not provide a stable, accuratemeasure of flow rate; thus, under these circumstances, the rotameters are relied on for flowrate measurements. At flow rates exceeding the combined full-scale reading of therotameters (24 + 60 = 84 gal/min), the turbine flowmeter provides an accurate means ofmonitoring flow rate through the check valve. Bypass flow rate was not measured.

The flow loop also includes a drain and cold water supply, the latter being a tanklocated above the ceiling of the room. Under normal flow operations, the loop watertemperature increases because of the energy supplied by the pump. Flow loop watertemperatures were stabilized by adjusting the water supply and drain flow rates so that thenecessary heat removal was achieved.

The acquired magnetic flux signatures (see Fig. 2.13) showed that at a low flowrate (insufficient to open the valve fully), the disc fluttered considerably in midstroke,whereas at a higher flow rate, the same valve achieved a fully open and stable condition.

21

ORNL DWG 90-3992 ETD

3-INCH SWING CHECK VALVE -DISC MOVED MANUALLY

zQU,

Z~O

ccw

0(nB

aw.

CU,>

dw1

NJ

TIME (s)

I MEASURED MECHANICALLY (ANGULAR DISPLACEMENT TRANSDUCER)2 GAUSSMETER PROBE INSTALLED ON VALVE CAP

Fig. 2.10.measurements.

Comparison of magnetic field strength and disc angular position

243 Detection of Worn Hinge Pins

Experiments carried out at ORNL have shown that MFSA techniques can be usedto detect hinge pin wear. Figure 2.14 illustrates a technique for detecting worn hinge pinsthat makes use of two Hall-effect gaussmeter probes, mounted so that each probe providesan independent measurement of instantaneous hinge arm position. When both probes aremounted on the valve cap at locations equidistant from and perpendicular to the projectedhinge arm travel plane, both gaussmeters should provide identical signatures when thehinge arm moves in a purely swinging motion as the valve opens and closes.

In addition to swinging, the hinge arm moves in a side-to-side rocking motion aswell, as a result of flow turbulence and the clearances between the hinge pin and hingearm. As this clearance increases (e.g., because of hinge pin wear), the propensity to rockincreases. Thus, the increase in hinge arm rocking is detected as increased deviations fromthe single line (pure swinging) relationship between the probe output signals, as shown inFig. 2.14.

22

ORNL PHOTO 8234-88

~Ii 1i

Fig. 211. Check valve [low loop (photo).

Another technique that appears to be useful for detecting worn hinge pins is basedon an analysis of the magnetic flux time waveform (signature) acquired by a single probeduring a full valve stroke. Figure 2.15 illustrates that, the time waveform changesappreciably when different hinge pins are installed. This reflects changes in hinge armposition due to differences in clearance between the hinge arm and hinge pin. During anopening or closing of the valve, the magnet (which is mounted on the hinge arm) rotatesand translates along a different path that is determined by this clearance.

Figure 2.15 shows that, even when the normal-sized hinge pin was installed, themagnetic field strength varies with valve position in a nonlinear manner. In fact, when thevalve is near the full-open position, the same magnetic field strength reading is reached attwo distinctly different valve positions. The relationship between the external magnetic fieldreading and valve position when the normal-sized hinge pin was installed is different thanthat previously discussed (see Figs. 2.10 and 2.13). This relationship was seen to beapproximately linear and without any "humps" in the time waveform that would result in twovalve positions existing for the same magnetic field reading. The differences in thesesignatures are simply a result of locating the gaussmeter at a slightly different position.

23

ORNL-DWG 90-3993

WATER SUPPLY

TURBINEFLOW

METERFLOW>-

ROTAMETERSBYPASS

GATEVALVE

24 GPM

BALLVALVE

FLOWCHECK VALVETEST SECTION

NEEDLEVALVE

PUMP

DRAIN

Fig. 2.12. Check valve flow loop (schematic).

During these tests, it was noted that the magnitude of the "hump" could also beaffected by the rate at which the valve opened and closed. For example, the faster thevalve opened, the smaller the hump became even though the full-open signal magnituderemained the same.

Thus, when using MFSA, the relationship between magnetic field strength readingsand disc position is determined from several factors, including the locations of the internallyinstalled magnet and the externally attached gaussmeter and the clearance between thehinge arm and hinge pin.

2.5 COMPARISON OF ACOUSTIC, ULTRASONIC, AND MAGNETIC MELTODS

The preceding three sections of this report have provided descriptions of threecheck valve monitoring methods that are useful in determining check valve position, motion,and leakage. These methods, based on acoustic emission, ultrasonic inspection, andmagnetic flux monitoring, function according to different principles of operation and thusprovide different (and complementary) diagnostic information. The currently estimatedcapability of each monitoring method to detect various check valve operational conditionsis given in Table 2.1.

J

24

ORNL-DWG 89.5104 ETD

0.5-- FLUTIER MAGNITUDE

o.4--

< 0 DISC FLUTTERING0

'O0.2

PUMP STARTED - LOW FLOW RATEI * (INSUFFICIENT TO OPEN VALVE FULLY)0w

0.5.DISC ON BACKSTOPUi 0.4-

U-0

p0.13z(9< 0.2

0.1

0 _0 PUMP STARTED - HIGH FLOW RATE0.0 (OPENS VALVE FULLY)

-0.1 I I ITIME (2 SECONDS/DIVISION)

Fig. 2.13. Usc of MFSA to detcct disc instability (fluttcr).


ci-E 0.40

; 0.30

X GAUSSMETER

Fig. 2.14. Dctccting worn hinge pins using MFSA.

25

ORNL-DWG 90-3394 ETM

Normal Hinge Pin

@2

0@2

I-

4)

@2

.2

C..

@2

C#

4)

S.-

C)

@2ID

4)

4A

S.-WU

4)

@2

0.40

0.30

0.20

0.10

0.00

-0.100.40

0.30

0.20

0.10

0.00

-0.100.40

0.30

0.20

0.10

0.00

-0.100 10 20 30 40 50 60 70 80 90 100 110. 120

Time (s)

Fig. 2.15. Magnetic flux time waveforms' of a 3-in, swing check valve for threehinge pin conditions: normal (top trace), medium worn (middle trace), and severely worn(bottom trace). All signatures were acquired using identical internal magnet and externalgaussmeter locations.

Table 2.1. Sclected diagnostic capabilities and limitations of three check valve monitoring method

Detects Detects Sensitivity Monitors discvalve Detects fluttering to position throughout Works

internal internal (no ambient the full range withMethod leakage impacts impacts) Nonintrusive conditionsb of disc travel all fluids

Acoustic Yes Yes No Yes Sensitive to No Yesemission externally

generatednoise/vibration

Ultrasonic No Yes Yes Yes Unknown Not in all Noinspection (indirectly) cases-because of

limited viewingangle of transducer

MFSA No Yes Yes No-requires Sensitive to Yes Yes(indirectly) initial nearby

installation of externalpermanent magnetic fieldsmagnet inside (e.g., fromthe valve motors)

aRadiography and pressure noise analysis methods are not summarized in this table. This table does not reflectother attributes such as cost, ease of use, etc. This table reflects the judgment of the author and is based on allinformation available at the time this report was prepared.

"Temperature and radiation effects are unknown.

a)C7%

27

As indicated in Table 2.1, although no single technique has the capability to detectall check valve operational conditions well, a combination of acoustic emission with eitherultrasonic inspection or MFSA can yield a monitoring system that succeeds in providingsensitivity to detect all major check valve operating conditions. Both acoustic/ultrasonic andacoustic/magnetic combinations have been tested.

The combination of acoustic emission and ultrasonic monitoring methods is describedlater in this report where commercial systems are discussed. The combination of acousticemission and magnetic flux monitoring methods was tested by ORNL on a check valvewhose disc was moved manually to simulate disc'fluttering at different disc positions.

As shown in Fig. 2.16, the acoustic signature did not provide direct indication ofdisc'position when the valve's disc was stationary in the fully open and fully closed positions,nor did it detect the slowly moving disc or disc lutter in midstroke. In all three tappingmodes (seat tapping, backstop tapping, and hinge arm rocking), the acoustic signaturedetected the tapping but not its location. The magnetic signature did not unambiguouslydetect the tapping but, in conjunction with the'acoustic signature, identified its location.


DISC MOVED MANUALLY

A VALVE OPENED SLOWLY8 MID-STROKE FLUTTERING

0.8 (HINGE ARM ROTATINGON HINGE PIN) D E

0.6 C. VALVE CLOSED SLOWLYD TAPPING ON SEAT

> 0.4 E TAPPING ON BACKSTOP,J UW F HINGE ARM ROCKING ON I

< 0.2 HINGE PIN I.

~~~~~- 8"fi c II Eo

)>1 0.0

-0.2

-0.6_ACSTIC SENSOR MOUNTED ON VALVE CAP

-0.8

0.4-

;- TIMEsOPEN0.3.

CLi W

* . -'., -.A

-~~~~'l 02. 116an11x1n cutc'intrs o hckvleude eea

-J wa

~0.0 CLOSED

-aj10 5 10 Is 20 25 30 35 40

* ~~~~~~~~TIME (s)

Fig.- 2.16. Magnetic flux -and acoustic 'signatures for a check valve under severalsimulated operational conditions.

28

2.6 OTHER METHODS

2.6.1 Radiography

The use of conventional and new Thigh-energy" radiographic techniques formonitoring the condition of check valves is described in Ref. 2. The following is a summaryof the information presented in that reference.

Conventional radiography has been successfully used by several nuclear utilities todetect certain types of failures and degradation of check valve internals. Examples of swingcheck valve failures that were detected with conventional radiography include missing valvediscs, locking devices, and nuts; broken disc studs; discs stuck in full-open position; etc. Inaddition, the integrity of discs and guide pins in globe-type stop-check valves has beenverified with this technique. It was pointed out that these cases involved relatively small(6-in. or smaller) valves and that when applied to larger valve sizes, the interpretation ofthe radiographic film becomes more difficult (because of decreased resolution) and thus maynot provide reliable results.

The major drawbacks associated with conventional radiography, identified in Ref. 2,include

1. Inability to radiographically inspect steel sections in the field that are thicker thanapproximately 5 in.

2. Excessive exposure time (and outage time) for field radiography.3. Inability to acquire good radiography quality in field practice under adverse plant

environmental conditions, particularly those of high temperature and background ionizingradiation typical of nuclear generating units.

High-energy radiographic equipment has recently been developed which overcomesthe limitations of conventional radiographic techniques, and it has been used by a nuclearutility to determine the integrity of swing check valve internals. This new technology isdescribed later in this report where commercial diagnostic systems are discussed.

2.6.2 Pressure Noise Monitoring

Early in the Phase II study, pressure noise (e.g., fluid pressure perturbations) wasidentified as a measurable parameter that might be useful in check valve diagnosticapplications. Thus, an evaluation of pressure noise signature analysis was carried out byORNL using a 2-in. Stockham swing check valve installed in a cold water flow loop locatedin Oak Ridge.

Pressure noise was sensed by two piezoelectric pressure probes, one locatedupstream and one located downstream of the check valve. Figure 2.17 illustrates a typicalinstallation of the two probes and gives a list of selected probe specifications. Throughoutthe flow tests, the distance from each probe to the check valve was varied, as well as thedepth of insertion into the flow stream.

Early test results indicated that pressure noise spectral characteristics were noticeablyaffected by flow rate. Figure 2.18 illustrates typical noise spectral signatures of both probesfor two flow rates, 20 and 40 gal/min, with both probes located 10 pipe diameters from thecheck valve centerline. Among the features seen were relatively sharp frequency peaksrepresenting pressure perturbations at the pump shaft frequency (motor speed) and at thepump vane pass frequency.

29


PROBE A CHECK VALVE PROBE B

I~ - I'I I| k_5D__1 10D 1'. 10D

PIEZOELECTRIC PRESSURE PROBES

PCB 102A02SENSITIVITY 50 MV/PSI

RESOLUTION 0.002 PSI

RESONANCE ,250 KHZ

RISE TIME 2 MICRO SEC

LINEARITY IV

probe specifications and points o.f installation.Fig. 2.17. Pressure

FLOW THRU LARGE ROTAMETERNB G ODB WTG H A 1.

0,

6

0.


RMS

i PROBE A

)IGP 10D

RMS

GPM PROBE B

-ODUPSTREAM

0 LIN X HZ 40C

FREQUENCY (HZ)

Fig. 2.1& Effect of flow rate on pressure noise spectra.

30

In addition to these peaks, the pressure noise frequency content of both probesignals, especially above 200 Hz, varied in amplitude (by up to a factor of 10) as a resultof changes in flow rate. Since the check valve disc position also changed with flow rate,it seemed likely that certain pressure noise signature features (i.e., those affected by flowrate) might directly reflect check valve disc position. However, it was soon discovered thatsignature reproducibility was not very good because of high sensitivity to many factors,including

1. Flow path.2. Water temperature.3. Tank water level (system pressure).4. Pressure perturbations induced by pipe vibrations.5. Air in the system-especially that which collected in the check valve bonnet.

For example, Fig. 2.19 illustrates the effect of flow path on pressure noise spectra. At thesame flow rate (60 gal/min), the pressure noise spectral characteristics changed noticeablyas a result of varying the flow path from (a) through the large rotameter to (b) throughthe ball valve.

Pressure noise spectra thus were found to be complex signatures that were greatlyinfluenced by many system parameters. Consequently, a large effort was made tounderstand (and eliminate, if possible) many of these effects. For example, tests were


60 GPMLOG A NB G 00B WTG H A 1.O V RMS

5 I 01303 1 1 I I

EU _ .__ TI-RU BALL VALVE- 1

tZ DOWNSTREAM

5E--

11-RU BALL VALVEJ

tU

0 LOG B EXPO N 64 9 2.O V RMSz s

TI- 1RU BALL VALVE - PROBE B

W EU 100UPSTREAM

SE-3 - e i TH-U ROTAMETER _

00 LIN X HZ 400

FREQUENCY (HZ)

Fig. 2.19. Effect of flow path on pressure noise spectra.

7 l

31

carried out at a constant loop water temperature. Care was taken in establishingreproducible tank water levels 'and flow paths. Pipe 'vibrations were monitored byaccelerometers, and the spectra were compared with the pressure noise spectra.

In addition, various signal analysis techniques were explored in an attempt todevelop a means of arriving at a diagnostic signature that provided maximum sensitivity tocheck valve effects while having minimum sensitivity to other influences. Those techniquesincluded determining

* the pressure probe signal ratio (PA/PB), where PA is the pressure signal from probe "A"(downstream from check valve) and PB is the pressure signal from probe "B" (upstreamfrom check valve);

* the pressure probe signal difference (PA - P8 ); and* the correlation between pressure probe signals.

These methods, while initially promising, also suffered from nonreproducible results.In conclusion, pressure noise spectra were complex signatures that were highly

sensitive 'to many system" effects and apparently to other (unidentified) effects as well.Pressure noise signature analysis thus was judged to be an unattractive monitoring methodfor check valve diagnostics and was not considered for further studies, primarily because ofa lack of reproducible results.

2.7 CONCLUSIONS-BENEFrS AND WEABKESSES OFMONITORINGMEITHODS

This chapter has described several check valve monitoring methods and hasidentified their strengths and weaknesses. Those methods are acoustic emission, ultrasonicinspection, magnetic flux monitoring, radiography, and pressure noise monitoring. Of thosemethods, the acoustic emission, ultrasonic inspection, and magnetic flux monitoring methodsprovide the greatest overall diagnostic capabilities.

Radiography certainly provides unique information in the form of images that canbe very useful in verifying the integrity of check valve internals or in detecting failures ordegradation. Based on information available at the time this report was prepared, its majordrawback is the lack of a high-resolution display that provides quantitative information onthe position of the check valve internals, especially over a short time frame or during atransient (opening or closing of the valve).

Pressure noise monitoring is intrusive-a transducer must be installed whichpenetrates a pressure boundary (e.g., pipe). In addition, pressure noise monitoring wasshown (on a laboratory scale) to be overly sensitive to extraneous inputs (flow rate, flowpath, fluid temperature, pipe vibrations, etc.).

The main limitation of acoustic emission is that the absence of detectable tappingnoise does not by itself guarantee that the check valve is fully open and stable since thedisc may be oscillating in midstroke without tapping, stuck in midstroke, or detached fromthe hinge arm and lying still in the bottom of the valve. A minor limitation of this methodis the necessity of using multiple sensors to determine the location of a tapping event.

Ultrasonic inspection, using a single transducer installed at a fixed position, may notprovide valve disc position information over the full travel of the disc because of the limitedviewing angle of the transducer. Furthermore, a low-density fluid, such as steam, may resultin severe attenuation of transmitted and reflected signals and, ultimately, poor transducerresponse.

32

It is generally more difficult to apply ultrasonic inspection techniques on stainlesssteel valves than on carbon steel valves since ultrasonic signals are attenuated (scattered)more when passing through stainless steel than when passing through carbon steel. Thick-wall stainless steel valves (e.g., 1500-psi service) present the greatest difficulty; however,additional research needs to be carried out in order to fully explore these limitations.

MFSA requires the installation of a permanent magnet inside the valve; thus, themethod is intrusive. Impacts between the valve disc and valve body may result in time ina demagnetization of the attached magnet. Furthermore, if the magnet (and/or magnetassembly) detaches from the check valve, it may present a serious problem. The internalmagnet may attract and hold small metallic particles. The particles can build up and canaffect the magnetic field dispersion pattern, thus possibly changing the strength of themeasured external field. Metallic particles may also accumulate around the internal magnetin a manner which could affect the operation of the check valve. Finally, the magnet fluxsignature features may be difficult to observe under field conditions because of the presenceof nearby ambient magnetic fields.

A check valve methodology that combines acoustic emission monitoring with eitherultrasonic inspection or magnetic flux monitoring provides the following general capabilities:

1. Detecting leakage through a closed valve. (Acoustics)2. Detecting impacts occurring within the check valve during flow operations. (Any

method)3. Detecting disc position at all times, whether the valve is experiencing full flow, partial

flow, or no flow. (Ultrasonics or magnetics)

When monitoring check valves using either of these combinations, it is not necessary toapply both methods simultaneously in order to achieve the combined capabilities of the twomethods. Two check valve diagnostic systems that are based on the combinations describedabove are commercially available. These and other commercially available systems aredescribed in the next chapter of this report.

33

3. DESCRIMPON OF SELECTED COMMERCLALLY AVAILABLECHECK VALVE DIAGNOSTIC SYSTEMS

In order to adequately describe the commercially available check valve monitoringsystems, information was requested from several vendors of diagnostic systems, includingthose discussed in this chapter. The descriptions provided below are based on theinformation provided by the vendors.

3.1 CHECKMATu II (HfNZE-MOVATS, INC)

At present, CHECKMATE7' II is the only commercially available check valvemonitoring system that is based on ultrasonic inspection. The system is available fromHenze-Movats, Inc., of Kennesaw, Georgia. This system represents an upgrade of theoriginal CHECKMATEV system. According to the vendor, the new system utilizesimproved hardware and software that provide a means of more easily acquiring andanalyzing check valve signatures. Figure 3.1 provides a simplified drawing that illustratesthe basic operation of the CHECKMATE' II system. One ultrasonic transducer is used(pulse-echo type) which provides both transmission and receiving (sensing) capabilities.*

This system utilizes signal processing circuitry that filters out undesirable ultrasonicsignal reflections present in the raw received signal so that the resultant processed signalprovides a more easily interpreted valve disc position signature. According to Henze-Movats, Inc., since March 1987, approximately 200 valves in 21 power plants have beentested using CHECKMATET' and CHECKMATEt' II systems.

According to Henze-Movats, Inc., in addition to disc position indication, theCHECKMATE' II data analysis program can also provide estimates of hinge pin wearrates and fatigue damage of valve internal parts.

Henze-Movats, Inc., has recently used acoustic emission monitoring with theirultrasonic inspection system. The result of using the combination of these methods is seenin Figs. 3.2 and 3.3,t which compare data from the two sensors obtained for a check valvetapping its backstop and its seat, respectively. In both cases, the acoustic signature detectedthe tapping but not its location. The ultrasonic signature did not clearly detect the tapping,but, in conjunction with the acoustic signature, identified its location.

32 QUICKCHECKu (LIBERTY TECHNOLOGY CENTER, INC.)

Another system that utilizes a combination of monitoring methods isQUICKCHECK10, a check valve diagnostic system available from Liberty Technology

*Letter from J. N. Nadeau, Henze-Movats, Inc., to H. D. Haynes, ORNL, Subject:Motor-Operated Valve and Check Valve Systems Information, dated March 1, 1990.

tLetter from D. M. Ciesielski, Henze-Movats, Inc., to H. D. Haynes, ORNL,Subject: CHECKMATE' System Information, dated July 22, 1989.

34

ORNL-DWG 90-3539 ETD\ULTRASONIC TRANSDUCER

\ W ~ULTRASON11INA COMPUTER-BASEDGENERATOR/ANALYZER SIGNAL ANALYZER

F:g. 31Asmpieddpcc H% L,

Fig. 3.1. A simplified depiction of the CHECKMA1IV II system.

ORNL-DWG 90.4000 ETD

CHECKIMATEV2ZT4

-0.71

BACKSTOP TAPPING BACKSTOP TAPPING BACKSTOP TAPPING

14.68

ACOUSTICVZT3

0.0 0.4 zoTIME (s)

2.8 3.6

Fig. 3.2 CHECKMATFV" and acoustic signatures for a check valve undergoingbackstop tapping. (Used with the permission of Henze-Movats, Inc.)

35

ORNL.DWG 90-4001 ETD

0.38

SEAT TAPPING

CHECK MATEV2ZTS

'PING

ACOUSTICV2%T7

12.53

_~ i___I.-_J 4.18

0.00

.4.,0.0 0.4 0.8 1.2 1.6 2.0

TIME (s)2.4 2.8 3.2 3.6 4.0

Fig. 3.3. CHECKMATED' and acoustic signatures for a check valve undergoingscat tapping. (Used with the' permission of Henze-Movats, Inc.)

Center, Inc., of Conshohocken, Pennsylvania.* QUICKCHECK"', depicted in simple formin Fig. 3.4, utilizes a combined acoustic/magnetic dual sensor to monitor simultaneously thestructurally transmitted acoustic noise that results from flow and internal part impacts andthe position and motion of an encapsulated magnet that is permanently installed on a checkvalve internal part (e.g., hinge arm, disc, etc.). Data acquisition hardware includes the dualsensor(s), signal conditioning electronics, and a digital audio tape recorder. Recordedsignals are then processed, displayed, and analyzed with a computer-based system thatprovides detailed analysis capabilities for both acoustic and magnetic signals.

3.3 VIP (CANUS CORPORATION)

CANUS Corporation of Laguna Hills, California, has developed a check valvediagnostic system called Valve Inspection Program (VIP). VIP utilizes accelerometers (from2 to 12) and associated electronics (charge amplifiers, signal conditioning) to obtain checkvalve acoustic emission information that is stored on a digital audio tape recorder forsubsequent manual.and automated analyses.t .,Real-time, .high-resolution displays of theacoustic waveforms are provided by a digital graphic recorder. In addition, headphones areused to monitor the signals for qualitative assessments of fow noise, valve tapping, etc.

*Letter' from D. Manin, Liberty Technology Center, to H. D. Haynes, ORNI,Subject: Motor-Operated Valve, and Check Valve Systems Information, dated March 1,1990.

tLetter from Peter Pomaranski, CANUS Corporation, to H. D. Haynes, ORNL,Subject: Valve Inspection Program Information, dated March 1, 1990.

36


DUAL SENSOR.(ACOUSTICS/MAGNETICS)

Fig. 3.4. Simplified depiction of the QUICKCHECK" system.

According to CANUS, over the last 3 years, they have used this method to testapproximately 150 valves at five plants. Their tests have provided indications of check valvedegradation and check valve operating modes that would lead to degradation.

CANUS is presently developing neural network programs that could provideautomated interpretation of acoustic emission signals.

3.4 AVLD (LEAK DETECTION SERVICES, INC)

Leak Detection Services, Inc., of Annapolis, Maryland, has developed an acousticvalve leak detector for use aboard U.S. Navy submarines; it has also been used to detectinternal valve leakage at several commercial nuclear and fossil power plants. The devicepermits the operator to observe the acoustic emission signals on a meter and to recordthem on an x-y plotter. The device provides the capability for acquiring -acoustic signalsfrom two sensors simultaneously, one sensor mounted on the valve and the other mountedon the pipe about 10 pipe diameters away from the valve.' The two channel responses arethen adjusted, and the background noise signal (acquired by the pipe-mounted sensor) iselectronically subtracted from the valve-mounted signal. A positive signature is anindication of valve leakage.

37

35 MINAC-6 (SCHONBERG RADIATION CORPORATION)

The MINAC-6 portable linear accelerator system, available from SchonbergRadiation Corporation of Santa Clara, California, provides several advantages overconventional industrial radiography systems.* It is a portable, high-energy (6-MeV,300-rads/min at 1 m) system that is capable of producing radiographic images through a12-in. section of steel with a 10-min exposure time and through a 14-in. section in 45 min.The MINAC system was developed in cooperation with the Electric Power ResearchInstitute and, since 1981, has been used at approximately 20 nuclear power plants in theUnited States for inspection of heavy wall components, including large piping, valves, andpumps. In particular, the system was used to verify the integrity of internal parts of severalmain steam check valves at the Diablo Canyon Power Plant (Unit 2) while the unit was on-line.

3.6 PHYSICAL ACOUSTICS CORPORATION

Physical Acoustics Corporation (PAC), of Princeton, New Jersey, manufacturesacoustic leak detection instruments as well as provides field test services for leak detectionsurveys. PAC has developed a dual-channel acoustic valve leak detector that has beenused by utilities to detect valve leakage for more than 10 years.t The detector usesacoustic sensors attached to the valve body to determine the acoustic signal levels as wellas frequency content. The output is recorded on an x-y plotter showing signal amplitudeand frequency.

*Product catalogs, "MINAC 1.5, 4/6, Portable Linear Accelerator Systems,"Schonberg Radiation Corporation, Santa Clara, CA.

tLetter from Sam Ternowchek, Physical Acoustics Corporation, to H. D. Haynes,ORNL, Subject: PAC Leak Monitors, dated December 14, 1990.

39

4. NUCLEAR INDUSTRY CHECK VALVE GROUP TESTOF CHECK VALVE DIAGNOSTIC SYSTEMS

4.1 GENERAL INFORMATION

Three commercial suppliers of check valve monitoring equipment recentlyparticipated in a comprehensive series of tests designed to evaluate the capability of eachmonitoring technology to detect the position, motion, and wear of check valve internals(e.g., disc, hinge arm, etc.) and valve seat leakage. Those vendors were Henze-Movats, Inc.(ultrasonic/acoustic emission); Liberty Technology Center, Inc. (magnetic/acoustic emission);and CANUS Corporation (acoustic emission). Those tests, which began in late January1990 and were completed in mid-March 1990, were directed by the Nuclear Industry CheckValve Group (NIC) and were carried out at the Utah Water Research Laboratory locatedon the Utah State University campus.

This chapter presents a brief discussion of these tests, including descriptions of thecheck valves used in the tests, the test conditions used, and selected vendor data. It shouldbe noted that this discussion is based on a limited amount of information that was obtainedprimarily from the diagnostic system vendors directly. The results from these tests will bedescribed in a final report that will be distributed to the NIC utility members thatparticipated in the funding of the activity. Because of the heavy involvement of vendorsin those tests, a comprehensive review and assessment of the NIC test methodologies,vendor test data, and NIC conclusions should be carried out by NRC after NIC issues theirfinal report.

Eleven check valves were used in the tests; Table 4.1 indicates their size,manufacturer, and type.

These tests were carried out using water as the process fluid. Check valves wereinitially tested in "new" condition and then with one or more simulated degradations and/oroperational failures. Several flow conditions (resulting in several check valve operationalmodes) were also used. The check valve degradations used in the tests are listed in Table4.2. Flow conditions and valve operational modes are indicated in Table 4.3.

It is recognized that data analysis and nonintrusive testing technology for checkvalves are evolving. After the NIC test results are made available to NRC, the results ofthese tests should be reviewed; if necessary, an updated report will be issued.

4.2 SELECTED TEST RESULTS

Figure 4.1 illustrates acoustic and magnetic signatures obtained by theQUICKCHECK"' system for a 12-in. tilting-disc check valve in 'new" condition.* Theacoustic trace contains a transient (spike) indicative of the impact that occurred when flow(through the valve) was shut off and the valve closed (seated). The magnetic trace shows

*Letter from D. Manin, Liberty Technology Center, to H. D. Haynes, ORNL,Subject: Motor-Operated Valve and Check Valve Systems Information, dated March 1,1990.

40

Table 4.1. Check valves tested during the NIC monitoring methodevaluation tests

Size (in.)

10

12

4

10

10

16

24

24

20

6

6

Manufacturer

Crane

Val-Matic

Velan

Mission

Velan

Rockwell

Val-Matic

Atwood & Morrill

Atwood & Morrill

Powell

Crane

Check valve type

Swing check

Tilting disc

Swing check

Duo-check (splitdisc)

Swing check

Tilting disc

Tilting disc

Swing check

Swing check

Swing check

Tilting disc

Body material

Carbon steel

Carbon steel

Carbon steel

Carbon steel

Carbon steel

Carbon steel

Carbon steel

Carbon steel

Carbon steel

Stainless steel

Carbon steel

Table 4.1 Check valve degradationsused in the NIC tests

* Induced seat leakage

* 15% and 30% hinge pin diameter reduction

* 15% and 30% disc stud diameter reduction

* Broken spring (for valves having springs)

* Disc stuck open

* Combinations of the above degradations

e Missing disc

41

Table 43. Flow conditions and valve operationalmodes used in the NIC tests

0

0

0

0

0

Zero flow

Backstop tapping with and without cavitation

Two mid-stroke flow conditions

Seat tapping

Simulated pump start and trip

Reverse flow

Maximum flow

Minimum flow to open


U)(ZC,)

C

-j

z

0

0

0

I -IMPACT WHEN_ Jo~~." VALVE CLOSES

. n~~~~~~~~~~~~~~

d

31 II _ .. - ~ _.. ,,> 1 I-

I I I

a6,

75% FLOW,c-

.' .(A NO HINGEWEAR

z

0U=

VALVE SEATEDz . BIVALVE LEAVES(D - -SEAT

TIME

Fig. 4.1. QUICKCHECKm' signals for a 12-in. tilting-disc check valve in 'new"3

condition. (Used with the permission of Liberty Technology Center, Inc.)

42

direct indication of disc travel in the closing direction prior to the detected impact. Thetwo signals together thus confirm that the valve closed. When flow was restored throughthe valve, the valve opened, as indicated by the magnetic trace. According to LibertyTechnology Center, Inc., the decrease in the magnetic signal magnitude, observed when thevalve initially lifts off its seat, indicates that the valve had been fully seated prior toopening.

Figure 4.2 shows acoustic and magnetic traces for the same valve, but using a hingepin with a 30% reduction in diameter. Changes in features were observed in both signals,including the absence of a momentary decrease in magnetic signal magnitude as the valveopened. This, according to Liberty Technology Center, Inc., indicates that the valve's discdid not fully seat; instead, it hung down past the seat because of the smaller hinge pin. Asthe valve opened, acoustic impacts were recorded which were a result of the increasedclearances between the hinge arm and the hinge pin. In addition, when the smaller hingepin was used, disc flutter was observed to occur when the valve reached its open position,as shown by the magnetic trace.

ORNL-DWG 90-4003 ETDC)I

<

o 0 4 LARGER IMPACT DUEcD TO 30% HINGE PIN WEAR

,_- __L____-______C)

a)

C,; A ,l

J [ ~~~~~~LARGER FLUTTER DUE

0

C' _ |___ TO 30% HINGE PIN WEARo ~~~~~~~VALVE NEVER SEATS

I- - DISK HANGS DOWN TOO FAR

w

z [/_______ LTHEREFORE, DISK IS NOT< ABLE TO SEAT

Fig. 4.2. QUICKCHECIC2 signals for a 12-in. tilting-disc check valve with 30%hinge pin diameter reduction. (Used with the permission of Liberty Technology Center,Inc.)

43

Figures 4.3 and 4.4 show the differences in the valve seating acoustic signatures thatoccurred as a result of the different hinge pin sizes. Figure 4.3 shows that a single acoustictransient occurred during disc closure when a normal hinge pin was used. Figure 4.4 showsthat, when the smaller hinge pin was installed, multiple impacts (resulting in several acousticsignal transients) occurred.

These examples further illustrate the complementary information that can beobtained from using both a disc position (in this case magnetic) monitoring method and animpact (acoustic emission) monitoring method.

The following two examples also illustrate the benefit of acquiring diagnostic datasimultaneously from multiple sensors of the same type-in this case, accelerometers,mounted at different locations. CANUS Corporation used several accelerometers installedon the body of each test valve at several locations, including

* Left and right side of the hinge pin.* Left and right side of the open stop position (either on the body or the bonnet,

depending on valve type).* Valve seat.


C.)

* C)_ _ , I

F~~~~~~I I_ ..

CZ

CD,C:,

, ME' I.1

-.

z

M ~ ~ ~ ~ ~ TM

Fig. 4.3. QUICKCHECI(" signals for a 12-in. tilting-disc check valve in 'newtcondition-cxpandcd scale to show seating transienL' (Used with the permission of LibertyTechnology Center, Inc.)

44

CD

0)

C

-j

z

0

00

'a

0

V)

CD

z

0

z


IFig. 4.4. QUICKCIIECK? signals for a 12-in, tilting-disc check valve with 30%

hinge pin diameter reduction-expanded scale to show seating transient (Used with thepermission of Liberty Technology Center, Inc.)

Figure 4.5 shows traces that were obtained from four accelerometers during one testcarried out on a 24-in. tilting-disc check valve.* This test was performed under full-flowconditions accompanied by induced flow turbulence. As shown in the figure, a hard impactwas detected by the accelerometer mounted on the left side of the hinge pin. A similarimpact was not indicated by the accelerometer mounted on the right side of the hinge pin.Thus, CANUS' interpretation was that the hinge pin was worn. The hinge pin was, in fact,degraded 15% during this test, verifying CANUS' prediction.

Figure 4.6 shows traces from the same four accelerometers, mounted on the samevalve, but now operating under full-flow conditions with only a small level of turbulence.

*Peter Pomaranski, CANUS Corporation, "Preliminary Overview of Findings andResults on Acoustic Emission Nonintrusive Check Valve Testing-Performed at the UtahWater Research Laboratory, Logan, Utah," presented at the Nuclear Industry Check ValveGroup Spring Meeting, April 24, 1990.


Series No. 7c Test No.410Trace #2

100% w/Turbulence Expanded @ 10KHZDate: 2/27190 Time: DAT 2:07Sampfe: 0.7 sec.

; _ | i , : ! .- ' . .; i : ,| ! . . ! ! * ' !

I ; * ; * . - ! r i i * . j ! iOPEN STOP RIGHT

.; , ., .HINGE PIN RIGHT _ i S I .. i i . ., , * ; .

v ! ! . ~~TAP Z..6Ai9

____ _____ ML

;HNGEPINRGHT -T -.::z~ -___ ,HINGE6PIN RIGHT

- l i, ! I . ! ' I ! ! X tHARD IMPACT ;HINGE PiN LEF .,-

F.g. 45. VIP (CANUS CoNporation) aGeleromeler traces from test 410 (2PNinHtilting-ic check Vavc, full-flow conditions, induced flo turbulence, worn hingc pin).Source: Peter Pomaranshi, CANUS Corporation, "Preliminary, Overview of Findings andResults on Acoustic Emission Nonintrusive Check Valve Testing-Performed at the UtahWater Research Laboratory, Logan, Utah," presented at the Nuclear Industry Check ValveGroup Spring Meeting, April 24, 1990.

ORNL-DWG904007 ETDSeries No. 7d Test No.420

Tape W2Trace U2100% w/Small Turbulence Expanded @ 10KHZ

Date: 2/27/90 Time: DAT 8:15:17 Sample: 0.7 sec.

;_i_._!_iii___:_!;_i___i!_Ii_.___j_*_ I - ;!!!i;*j

_______________________ _=}_____ ; III I--- OPEN STOP RIGHT-

OPEN STOP LE, I i ? i_? 'N i iLEFI ! _ _ _ _ -, -, - - - !,F ! ! !; i; ! H I N G E P IN i _ _ _ t _ ! _ ! _

=,.~~~1TiririiT} -r 1,_--~~

1 _

HE PiN LEFTt III IHI "~ _ HINGE PIN LEFT __ __ _ _ _ _ _ __ _ _ _ _

0o

Fig. 4.6. VIP (CANUS Corporation) accclcrometcr traces from test 420 (24-in.tilting-disc check valve, full-flow conditions, minimal flow turbulence, wom hingc pin).Source: Pctcr Pomaranski, CANUS Corporation, "Preliminary Overview of Findings andResults on Acoustic Emission Nonintrusive Check Valve Testing-Performed at the UtahWater Research Laboratory, Logan, Utah," presented at the Nuclear Industry Check ValveGroup Spring Meeting, April 24, 1990.

47

The decreased flow turbulence can clearly be seen by comparing the reduced signal noiselevel in these traces with the much noisier traces shown in Fig. 4.5. This figure also showsan impact that was detected by all accelerometers; however, the right-side hinge pinaccelerator detected an erratic ringing pattern rather than the clear ringing pattern indicatedby the left-side hinge pin accelerometer. This feature also led CANUS to predict that thehinge pin was worn, which, in fact, it was.

Figure 4.7 shows ultrasonic signatures, obtained by the CHECKMATEth II systemduring the NIC tests, for the same valve under two different flow conditions. As shown inthe figure, the degree of disc flutter varies with the flow rate, with the largest flutteroccurring at 1295 gal/min. Disc flutter is quantified in both plots as a measure of the discangular movement per unit of time (e.g., at 2251 gal/min, the flutter is 2.83 degrees/s,whereas at 1295 gal/min, the flutter is 3.65 degrees/s. At 2251 gal/min, theCHECKMATE7' II signal magnitude occasionally reaches its maximum value (ofapproximately 14.35 in.), indicating that the valve is tapping its backstop. It is noted thatthe restricted movement of the disc (as a result of tapping) results in a lower overall fluttermagnitude. In this case, the ultrasonic transducer was located on the bottom of the valve;therefore, the largest signal was produced when the disc was at the open position (at theposition furthest from the transducer).

Figure 4.8 illustrates that CHECKMATE7' II can be used to track the motion ofa check valve disc (e.g., in this case, the disc of a 12-in. Val-Matic tilting-disc check valve)from the full-open to full-closed position. In contrast to Fig. 4.7, the ultrasonic transducerwas located on top of the valve; therefore, the largest signal was produced when the disc

ORNL-DWG 90008 ETD

CHECKMATE Software Signature Plot(s)

14.32- Valve Id

c) \ P \ IGroup Name

c 14.12- MN|254 R Z 28.4 30 KNGE

Trace Name

.U . : ~~~~~~~~2.83 deg/sec .,TW .P o W/7"ruesnm13.92

0.00 f.17 2.34 3.51 4.68 5.85 7.02 8.19 Secs

12.4a- .- . - . :- -; , . . A ~~Valve Id:

a) Group Name :c 12.13- F 2 XI HVM

3.65 deg/sec Trace Name :125. p 3/4 FuLO WrnM echo

11.85.0.00 1.17 2.34 3.51 4.68 5.85 7.02 8.19 Secs

Fig. 4.7. Two ultrasonic signatures, obtained using the CHECKMAT1' II systemfor the samc valve (10-in. Vlan swing check) under different flow conditions: 2251 gal/min(top trace) and 1295 gal/min (bottom trace). (Used with the permission of Henze-Movats,Inc.)

48

ORNL-DWG90-4009 ETD

CHECKMATE Software Signature Plot(s)

11.309 -

10.803 -

10.300

9.795 -

a,0U

C

A_.|.. .. . . . . = . ,

FULL CLOSED

-- -I , I .

, :1 ; ' ~~~~~~~~~~~~~~* I \

.-- I FUL OPEN\

Valve Id :

HIC YVAE *2 - W VAL-4TIC

Group Name :FM 96 BASBINE AUL O1WaE

Trace Name :

WE - MOSE - WM

_-

9.292 .

B.787

8. 284 4

7.778

7.2750 -0.00 1'. 23 2~.45 3'.68 4'.91 6'.14 7'.36 Secs

Fig. 4.8. A CHECKMATN' H signature for a 12-in. Val-Matic tilting-disc checkvalve) from the full-open to full-closed positions. (Used with the permission of Henze-Movats, Inc.)

was at the closed position. It is recognized, however, that a single ultrasonic transducer(installed at a fixed location) may not provide valve disc position information over the fulltravel of the disc of some valves because of valve geometry and the limited viewing angleof the transducer.

Finally, Fig. 4.9 illustrates that a stuck check valve disc can be detected simply bycomparing the CHECKMATETh II signature obtained at no flow to a CHECKMATEIII signature obtained at a significant flow rate.

49

ORNL-DWG90-4010 ETDCHECKMATE Software Signature PlotCs)

7.742 -

Ona)r 5.161 -UuI

2.581 -

O0.000 -_0.00

7.742 -

a)r 5.161 -CI-'

2.581

Valve Id :VALVE t2 VAL-*ATIC t2 TD.

Group Name :

run 128 stuck open tt hinge

Trace Name :

zero flow stuck open

o gcTfl

1'.23 2.45 3.b8 4.91 6.14 7.36 Secs

; . . . Valve Id:

VALVE 12 VL.-ATIC 12- T.t.

i z ' Group Namej run 13t stuck cGP" t55 Wlnge

Trace Namefull flow lilneck

* I ' 2392 gpm0.000 4-

00 1.'23 2'.45 3'.68 4'.91 6'.14 7'.36 Secs

Fig. 4.9. CIECKMATJ"' II signature for a 12-in. Val-Matic tilting-disc checkvalve having a stuck disc. The top trace was obtained at no-flow conditions, while thebottom trace was obtained at 2392 gal/min (more than would be sufficient to move thedisc). (Used with the permission of Henze-Movats, Inc.)

I

51

5. SUMMARY AND RECOMMENDATIONS

The primary objective of the NPAR Program Phase II check valve aging assessmentis to identify and recommend methods of inspection, surveillance, and monitoring that willprovide timely detection of check valve degradation and service wear (aging) so thatmaintenance or replacement can be performed prior to loss of safety function(s). Insupport of the NPAR Program, ORNL has evaluated several developmental and/orcommercially available check valve diagnostic monitoring methods. In each case, theevaluations have been focused on the capability of each method to provide diagnosticinformation useful in determining check valve aging and service wear effects (degradation),check valve failures, and undesirable operating modes. Of those methods examined,acoustic emission monitoring, ultrasonic inspection, and magnetic flux signature analysisprovided the greatest level of diagnostic information. These three methods were shown tobe useful in determining check valve condition (e.g., disc position, disc motion, and seatleakage), although none of the methods was, by itself, successful in monitoring all threecondition indicators. However, the combination of acoustic emission with either ultrasonicor magnetic flux monitoring yields a monitoring system that succeeds in providing thesensitivity to detect all major check valve operating conditions. When monitoring checkvalves using either of these combinations, it is not necessary to apply both methodssimultaneously in order to achieve the combined capabilities of the two methods.Commercial versions of all three methods are still under development, and all shouldimprove as a result of further testing and evaluation.

The NIC test produced a large volume of useful check valve diagnostic data fromthree commercial systems; however, no test data were available (from NIC) for our reviewprior to the preparation of this report. Because of the significance of the data obtainedand its impact on determining operational readiness of nuclear plant check valves, it isrecommended that a comprehensive review and assessment of the NIC testresults-including NIC test methodologies, vendor test data, and NIC conclusions-becarried out by NRC after NIC issues its final report. In addition, any issues not addressedadequately by the NIC tests and deemed important in regard to safety-related issues shouldbe considered for further study.

53

REFRENCES

1. U.S. Nuclear Regulatory Commission, Nuclear Plant Aging Research (NPAR) ProgramPlan: Components, Systems and Structures, USNRC Report NUREG-1144, Rev. 1,September 1987. Available for purchase from National Technical Information Service,Springfield, Virginia 22161.

2. MPR Associates, Inc., and Kalsi Engineering, Inc., Application Guidelines for CheckValves in Nuclear Power Plants, Electric Power Research Institute, Inc., Report NP-5479,January 1988. Available for purchase from Research Reports Center (RRC), Box50490, Palo Alto, California 94303.

3. W. L Greenstreet, G. A. Murphy, R. B. Gallaher, and D. M. Eissenberg, MartinMarietta Energy Systems, Inc., Oak Ridge Natd. Lab., Aging and Service Wear of CheckValves Used in Engineered Safety-Feature Systems of Nuclear Power Plants, USNRCReport NUREG/CR-4302, Vol. 1 (ORNL-6193/V1), December 1985. Available forpurchase from National Technical Information Service, Springfield, Virginia 22161.

4. U.S. Nuclear Regulatory Commission, Loss of Power and Water Hammer Event at SanOnofre, Unit, 1, on November 21, 1985, USNRC Report NUREG-1190, January 1986.Available from National Technical Information Service, Springfield, Virginia 22161.

5. Letter from Steven A. Varga, NRC, to All Holders of Light Water Reactor OperatingLicenses and Construction Permits, Subject: Guidance on Developing AcceptableInservice Testing Programs (Generic Letter No. 89-04), dated April 3, 1989. Availablein NRC Public Document Room for inspection and copying for a fee.

6. J. G. Dimmick and J. M. Cobb, "Ultrasonic Leak Detection Cuts Valve MaintenanceCosts," Power Engineering 9(8), 35 (August 1986). Available in public technical libraries.

55

Appendix A.

REGULATORY TEST REQUIREMESM AND CURRENT TEST METHODS

REGULATORY TEST REQUIREMTM

There are three distinct regulatory requirements related to check valves. These arethe in-service testing (IST) requirements, containment isolation valve leak test requirements,and reactor coolant pressure boundary isolation valve leak test requirements.

In-Service Testing Requirements

Plant Technical Specifications, which invoke the authority of 10 CFR Part 50.55a,require that pumps and valves be tested in accordance with Sect. XI of the ASME Code.The scope of valves to be included in the IST program is defined in Article IWV-1000 ofSect. XI to include valves", which are required to perform a specific function! in shuttingdown a reactor to the cold shutdown condition or in mitigating the consequences of anaccident.".

Valves are categorized by function and functional requirements in Article IWV-2000.All check valves fit in Category C.-!valves which are self-actuating in 'response to somesystem characteristic, such as pressure (relief valves) or flow direction (check valves)." Inaddition, some check valves may be defined as Category A valves-valves for which seatleakage is limited to a specific maximum amount in the closed position of fulfillment oftheir function."

Article IWV-3000 of Sect. XI provides test requirements for valves in general, andArticle IWV-3520 specifically identifies those requirements that are particular to checkvalves. In essence, check valves are required to be demonstrated to be able to move totheir safety-related position when system conditions so warrant. Testing is required to beperformed quarterly; however, valves that cannot be tested quarterly are to be tested atcold shutdown. Each plant must submit a test program to the NRC which designates valvesthat are to be tested in accordance with the Code and which identifies exceptions to Coderequirements.

Containment Isolation Valve Lcak Tcsting Requirements

Plant Technical Specifications also invoke the authority of 10 CFR 50, Appendix J,and require that containment isolation valves that are- subject to 'Type C" (not to beconfused with "Category C" valves identified in the ASME Code) tests be leak testedbiennially. The valves that are specified in Appendix J are ". . those that:

1. Provide a direct connection between the inside and outside atmospheres of the primaryreactor containment under normal operation, such as purge and ventilation, vacuumrelief, and instrument valves; -

2. Are required to close automatically upon receipt of a containment isolation signal inresponse to controls intended to effect containment isolation;

3. Are required to operate intermittently under postaccident conditions; and

56

4. Are in main steam and feedwater piping and other systems which penetrate containmentof direct-cycle boiling water power reactors."

The specific containment isolation valves that are actually "Type C" leak tested arenormally identified in the individual plant's Final Safety Analysis Report.

Reactor Coolant Pressure Boundary Isolation Valve Testing Requirements

Plant Technical Specifications require that certain valves that provide isolationbetween the reactor coolant system (RCS) and connected low-pressure systems (e.g., theresidual heat removal system) be leak tested periodically. The test intervals are variable,depending upon plant and valve operation conditions, but testing is required at least onceper refueling outage. The maximum allowable leakage rate for each RCS pressureboundary isolation valve is specifically designated in the Technical Specifications.

There is overlap between these three test requirement sources. For example, sincecontainment' isolation valves are needed in mitigating the consequences of an accident,they would fall under the auspices of both the 10 CFR 50, Appendix J, requirements- andthe IST requirements. In general, the IST requirements cover the broad range of checkvalves to be tested, while the containment isolation and RCS pressure boundary isolationvalve leak tests apply specific limits to select valves.

CURRENT TEST METHODS

Utilities write and implement procedures to meet the regulatory requirementsassociated with check valves. The procedure used to verify a particular valve's operabilitymay range from a simple test requiring little or no test setup or data analysis, to a complextest evolution involving system realignment, partial draining, application of special testingequipment, post-test filling, venting, and realignment, and considerable data analysis.

Depending upon what function is to be demonstrated (e.g., closing or opening ondemand or seat leakage less than allowable), a number of conditions are monitored toindicate valve functionability. These include, for example,

a. Measuring the seat leak rate.b. Observing upstream temperature (as is done routinely, for example, in the case of

auxiliary feedwater check valves).c. Verifying that the required flow rate can be passed through the valve.d. Observing reverse pressure differential.e. Listening for the valve'to slam shut on cessation of flow.f. Observing that the valve is sufficiently closed to ensure that the rate of flow delivered

downstream of the valve to a given target is adequate to meet system demands.

For check valves that are required to transfer from closed to open to satisfy theirsafety-related function, testing is conducted by passing the required flow rate through thevalve, when practicable. The NRC has historically accepted full-flow testing as evidencethat the valve has been full-stroked (it is recognized, however, that in some applications,valves may not fully open, even under maximum system flow). Valves that are required to

57

transfer from open to closed are normally tested by applying a reverse pressure differentialand observing some system parameter for indication of closure.

Typically, a relatively small fraction of the check valves used in safety-relatedapplications are leak tested to determine a quantitative leak rate. ORNL conducted asample review of the IST programs for four units. Included in the review were twopressurized water reactor units and two boiling water reactor units. There was a total of466 check valves at the four units. Only 118 of these valves were identified as ASMECategory A valves, that is, "valves for which seat leakage is limited to a specific maximumamount in the closed position of fulfillment of their function." All 118 valves that weredesignated to be seat leak tested are tested under the auspices of either the containmentisolation valve or the RCS pressure boundary isolation valve leak testing programs, with noadditional leak testing specifically for IST purposes. Quantitative seat leakage testing isconducted by pressurizing the downstream side of the valve with a test pressure source (e.g.,instrument air or a compatible water source, including, in some cases, the existingdownstream pressure) and observing the leak rate by opening an upstream connection.

It should be noted that the safety-related function of some valves that are notquantitatively leak tested involves only allowing forward flow; nevertheless, a significantnumber of valves for which the safety-related function includes closing in response toreverse differential pressure are not designated as requiring leak testing. This includesvalves such as safety-related pump discharge check valves.

Seat leakage for valves in applications where quantitative leak rate is not requiredto be measured but where the valve must still be demonstrated to close in response toreverse differential pressure is typically addressed in a qualitative manner. For example, thepressure differential across a pump discharge check valve may be observed while the pumpis idle and a parallel pump is being run (or the downstream section of the check valve'spiping is otherwise pressurized). Alternatively, the adequate functioning of a valve, in termsof seating, may be demonstrated by verifying that the required flow rate is delivered throughthe demand flow path (thereby proving that the valve in question provided sufficientisolation).

In the event that the required testing is not practicable during normal operation, thevalve may be partially tested during normal operation (e.g., by passing a flow rate that issomething less than that required to satisfy the safety function) and then tested at therequired flow rate during cold shutdown or refueling. It should be noted that some valvescannot be even partially tested during normal operation because of the adverse effect thattest conduct would have on the plant.

It is important to note that the current test requirements are not oriented towardtrending or detection of conditions that may lead to valve failure. Disk flutter due tooperation under turbulent conditions with the valve not in a fully open position (pinnedagainst the backstop, for example) has been recognized as a cause of wear at the hingepin pivot area. Substantial hinge pin or pin holder wear can occur without being detectedby current monitoring. Other degradation mechanisms, such as disk pin wear, may also notbe detected. Even catastrophic failure can go undetected for valves in certain applications(e.g., a check valve that is forward flow tested only may pass its flow test with the diskhaving broken off and fallen to the bottom of the valve body or traveled downstream). Inaddition to the inability to detect some degradation/failure sources by use of historicalmonitoring methods, it is important to recognize that in some system designs/applications,it is simply not feasible, because of inadequate testability provisions in the system design,to conduct testing at the required conditions at any time.

58

In recognition of the weaknesses in historical monitoring capabilities and the inabilityto test certain valves, utilities have undertaken, at the urging of both the NRC and INPO,periodic disassembly and inspection. Disassembly and inspection has normally been usedon a sample basis-that is, a representative valve of a group of valves of similardesign/application is disassembled during a refueling outage. During the following outage,another in the group is inspected, and so on, until the entire group has been inspected.Typically, group size and sampling rates are such that all valves in a group are inspectedwithin 4 to 6 years.

While disassembly and inspection provide better information about valve conditionin many respects than can be obtained through any other available method, there are anumber of drawbacks associated with disassembly. These include, for example, schedulingadditional maintenance work during already busy outages, additional radiological burden, aswell as concerns that reassembly errors can go undetected (for valves that cannot be testedwith flow).

The need to improve the knowledge of check valve operating conditions withoutrequiring disassembly has resulted in the development and improvement of severalnonintrusive diagnostic techniques. It is important to recognize, however, that for thoseapplications in which the utility has determined that current system configuration does notpermit valve stroking, neither the historical nor the newer techniques are able to confirmvalve operability.

59

NUREG/CR-4302Volume 2ORNL-6193/V2Dist. Category RV

Intemal Distribution

1. J. M. Asher 52. R. C. Kryter2. E. L. Biddle, Sr. 53. K H. Luk3. S. H. Buechler 54. J. C. Moyers4. D. A. Casada 55. G. A. Murphy5. D. F. Cox 56. C. E. Pugh

6-16. D. M. Eissenberg 57. ORNL Patent Office17. E. C. Fox 58. Central Research Library18. R. H. Greene 59. Document Reference Section

19-49. H. D. Haynes 60-61. Laboratory Records Dept.50. D. E. Hendrix 62. Laboratory Records (RC)51. J. E. Jones Jr.

External Distribution

63. G. Sliter, Electric Power Research Institute, P.O. 10412, Palo Alto, CA 9430364. J. W. Tills, Institute for Nuclear Power Operations, 1100 Circle 75 Parkway, Atlanta,

GA 30339-3064 -

65. J. H. Taylor, Brookhaven National Laboratory, Bldg. 130, Upton, NY 1197366. A. B. Johnson, Battelle-PNL, MS P8-10, P. 0. Box 999, Richland, WA 9935267. Brian Stewart, 557 South Walnut, Apt #8, Cookeville, TN 3850168. J. P. Vora, U.S. Nuclear Regulatory Commission, Office of Nuclear Regulatory

Research, Electrical and Mechanical Engineering Branch, 5650 Nicholson Lane,Rockville, MD 20852

69. G. H. Weidenhamer, U.S. Nuclear Regulatory Commission, Office of NuclearRegulatory Research, Electrical and Mechanical Engineering Branch, 5650 NicholsonLane, Rockville, MD 20852

70. E. J. Brown, U.S. Nuclear Regulatory Commission, Office for Analysis and Evaluationof Operational Data, Reactor Operations Analysis Branch, Maryland National BankBuilding, 7735 Old Georgetown Road, Bethesda, MD 20814

71. C. Michelson, Advisory Committee on Reactor Safeguards, 20 Argonne Plaza, Suite365, Oak Ridge, TN 37830

72. M. Vagins, U.S. Nuclear Regulatory Commission, Office of Nuclear RegulatoryResearch, Chief, Electrical and Mechanical Engineering Branch, Division ofEngineering, 5650 Nicholson Lane, Rockville, MD 20852

73. W. S. Farmer, U.S. Nuclear Regulatory Commission, Office of Nuclear RegulatoryResearch, Electrical and Mechnical Engineering Branch, 5650 Nicholson Lane,Rockville, MD 20852

74. M. J. Jacobus, Sandia National Laboratories, P.O. Box 5800, Division 6447,Albuquerque, NM 87185

60

75. H. L. Magleby, Idaho National Engineering Laboratory, MS 2406, P.O. Box 1625, IdahoFalls, ID 83415

76. Joseph N. Nadeau, HENZE-MOVATS, Inc., 200 Chastain Center Blvd., Kennesaw, GA30144

77. Richard H. Tuft, Liberty Technologies, Inc., Lee Park, 1100 E Hector Street,Conshohocken, PA 19428

78. Peter Pomaranski, Canus Corporation, 23030 Lake Forest Drive, Laguna Hills, CA92653

79. W. M. Suslick, Duke Power Company, McGuire Nuclear Station, P.O. Box 488,Cornelius, NC 28031

80. John W. McElroy, Philadelphia Electric Company, Testing & Laboratories Division,P.O. Box 650, Valley Forge, PA 19482

81. Richard J. Colsher, EPRI M&D Center, 3 Industrial Highway, Eddystone, PA 1901382. Ron D. Endicott, Public Service Electric and Gas, P.O. Box 570, Newark, NJ 0710183. Richard G. Simmons, Tennessee Valley Authority, 1101 Market Street, BR 5S73A-C,

Chattanooga, TN 30402-280184. Michael T. Robinson, Baltimore Gas & Electric Company, Routes 2 & 4, Lusby, MD

2065785. Bob Parry, Public Service of New Hampshire, P.O. Box 300, Seabrook Station,

Seabrook, NH 0387486. T. F. Hoyle, WPPSS Mail Drop 520, 3000 George Washington Way, P.O. Box 968,

Richland, WA 9935287. Office of Assistant Manager for Energy Research and Development, Department of

Energy, Oak Ridge Operations Office, Oak Ridge, TN 3783188-89. Office of Scientific and Technical Information, P.O. Box 62, Oak Ridge, TN 37831

90-419. Given distribution as shown in NRC category RV (10 - NTIS)

N2-29 3 U.S.NUCLEARREGULATORYCOMMISSION 1 REPORT NUMBERE5RCM910? [Assigned by NFRC, Add Vol., Swoot.. Rev..NRCM 1102 end Addendum Nuumbes. If any.)3201,3202 BIBLIOGRAPHIC DATA SHEET NUREG/CR-4302,Vol. 2

{See instructions on the reverse) ORNL-6193/V22.TITLE AND SUBTITLE

Aging and Service Wear Of Check Valves 3. DATE REPORT PUBLISHEDAging and Service Wear of Check Valves ~~MONTH EA

Used in Engineered Safety-Feature Systems April 1991of Nuclear Power Plants; Aging Assessments 4. FIN OR GRANT NUMBER

and Monitoring Method Evaluations B08285. AUTHOR(S) 6. TYPE OF REPORT

TopicalH. D. Haynes

7. PERIOD COVERED lsnclisse Digest

B. PER FORMING ORGAN IZAT ION -NAME AND ADD R ESS aIf NRC. porovr D;vis;on. Office or Region. U.S. Nuclear Regulatory Commission, and maing add'ess: i contractor. pror.iname and mailing adress.J

Oak Ridge National LaboratoryOak Ridge, TN 37831-6285

9. SPONSOR ING ORGANIZATION -NAME AND ADDRESS WIf NRC. type "Same as abo v,- if conurtcbor provide NRC ODeisson. Office or Region. U.S. Nuclea Regularorv Commission,and miling address.)

Division of EngineeringOffice of Nuclear Regulatory ResearchU. S. Nuclear Regulatory CommissionWashington, DC 20555

10. SUPPLEMENTARY NOTES

11. ABSTRACT (200 words or kse

Check valves are used extensively in nuclear power plant safety systems and balance-of-plant systems. Thefailures of these valves have 'resulted in significant maintenance efforts and, on occasion, have resulted in water hammer,overpressurization of low-pressure systems, and damage to flow system components. These failures have largely beenattributed to severe degradation of internal parts (e.g., hinge pins, hinge arms, discs, and disc nut pins) resulting frominstability (flutter) or check valve discs under normal plant operating conditions. Present surveillance requirements fornuclear power plant check valves have been inadequate for timely detection and trending of such degradation becauseneither the flutter nor the resulting wear can be detected prior to failure. Consequently, the U.S. Nuclear RegulatoryCommission has had a continuing strong interest in resolving check valve problems.

In support of the Nuclear Plant Aging Research Program, Oak Ridge National Laboratory has carried out anevaluation of several developmental and/or commercially available check valve diagnostic monitoring methods, inparticular, those based on measurements of acoustic emission, ultrasonics, and magnetic flux. In each case, theevaluations have been focused on the capability of each method to provide diagnostic information useful in determiningcheck valve aging and service wear effects (degradation), check valve failures, and undesirable operating modes.

A description of each monitoring method is provided in this report, including examples of test data acquiredunder controlled laboratory conditions. In some cases, field test data acquired in situ are also presented. The methodsare compared. and sueeested areas in need of further develonment are identified.

12. KEY WORDS/DESCR !PTORS IListv w'ords Of phrases thatessist researchers In locating the report.) 13. AVAILABILITY STAT EMENT

Unlimited1<. SECURITY CLASSIFICATION

(This Page)

Valves, aging, service wear, maintenance, degradation, trend Unclassifiedmeasurable parameter, incipient failure, inspection, surveillance, UThisse sfliedmonitoring, functional indicators, signature analysis, acoustic emission,ultrasonic inspection, magnetic flux 15. NUMBER OF PAGES

16. PRICE

WRC FORM 335 1249)

THIS DOCUMENT WAS PRINTED USING RECYCLED PAPER

UNITED STATESNUCLEAR REGULATORY COMMISSION

* WASHINGTON, D.C. 20555

OFFICIAi BUSINESSPENALTY FOR PRIVATE USE, $300

SPECIAL FOURTH-CLASS RATEPOSTAGE E FEES PAID. S. , USNRC

. PERMIT No. G-67

:

'V

C1201

Pi;4

�v2rT2

r

i

� ; - -.I

I I

I

- Aging- a:rd Service Wear of Check Valves Used---in Engineered · 2012. 11. 19. · NUREG/CR-4302...

Documents

Transcript of - Aging- a:rd Service Wear of Check Valves Used---in Engineered · 2012. 11. 19. · NUREG/CR-4302...